arenhoev@kph.uni-mainz.de
Hartmut Backe
backe@kph.uni-mainz.de
Dieter Drechsel
drechsel@kph.uni-mainz.de
Jorg Friedrich
friedrch@kph.uni-mainz.de
Karl-Heinz Kaiser
kaiser@kph.uni-mainz.de
Thomas Walcher
walcher@kph.uni-mainz.de
University of Mainz
Institute for Nuclear Physics
Johann-Joachim-Becher-Weg 45
55128 Mainz, Germany
The articles in this book originally appeared on the internet (www.eurphysj.org) as open access
publication of the journal
The European Physical Journal A Hadrons and Nuclei
Volume 28, Supplement 1
ISSN 1434-601X
c SIF and Springer-Verlag Berlin Heidelberg 2006
This work is subject to copyright. All rights reserved, whether the whole or part of the material is concerned,
specically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microlm or in any other way, and storage in data banks. Duplication of this publication or parts thereof
is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current
version, and permission for use must always be obtained from SIF and Springer. Violations are liable for
prosecution under the German Copyright Law.
c SIF and Springer-Verlag Berlin Heidelberg 2006
Printed in Italy
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specic statement, that such names are exempt from the relevant protective laws
and regulations and therefore free for general use.
Foreword 71 M. Vanderhaeghen
Two-photon physics
81 M. Ostrick
Many Body Structure of Strongly Electromagnetic form factors of the nucleon
Interacting Systems Experiments at MAMI
19 W.U. Boeglin
Few-nucleon systems at MAMI and beyond 107 F.E. Maas
Parity-violating electron scattering at the MAMI
facility in Mainz
29 D. Rohe The strangeness contribution to the form factors
A1 and A3 Collaboration of the nucleon
Experiments with polarized 3 He at MAMI
117 N. dHose
39 M. Schwamb Virtual Compton Scattering at MAMI
Few-nucleon systems (theory)
Foreword
This volume contains the proceedings of the Symposium on Twenty Years of Physics at the Mainz Microtron
(MAMI), which was held at the Johannes Gutenberg-Universitat Mainz, October 19-22, 2005. The Symposium marks
the retirement of several members of the Institut fur Kernphysik whose work has been devoted primaryly too scientic
research at MAMI over many years. It was the primary aim of the Symposium to review past and current activities in
the eld of hadronic structure investigations with the electroweak interaction. However, the Symposium also gave an
outlook on the physics with the MAMI upgrade, a double-sided mictrotron that is expected to provide a high-quality
beam of up to 1.5 GeV later this year.
The Institut fur Kernphysik was founded in the early 1960s by the late Hans Ehrenberg who served as its director
for more than two decades. He provided the Institute with a 350 MeV pulsed linear electron accelerator, which became
available in 1966 for studies of charge and magnetization distributions in nuclei and nucleons as well as photonuclear
investigations in collaboration with the Max-Planck-Institut fur Chemie.
Hans Ehrenberg knew about the importance of having excellent facilities for performing outstanding physics from
his earlier studies at Bonn and Stanford, with the later Noble Prize winners Paul and Hofstadter, respectively. There-
fore, he dedicated great eort in I) building up a perfect infrastructure of mechanics, electronics, vacuum and computer
workshops, and II) attracting a young accelerator physicist, Helmut Herminghaus, to the Institute.
In the late 1960s it became common wisdom that the next accelerator generation had to provide a high duty-
factor in order to perform coincidence experiments for detailed studies of hadronic physics. Helmut Herminghaus had
conceived a blueprint for such a device in 1975, a three-stage racetrack microtron (RTM). Shortly after a physics
program around this RTM was worked out and the proposal was sent to the sponsoring agencies. The project received
the support of the University and the State of Rheinland-Pfalz and sometime later also of the federal agencies. In
the fall of 1978, the state minister was informed by the federal minister of research and technology (BMFT) that the
project had been discussed with the German Science Council, the Deutsche Forschungsgemeinschaft (DFG), the Max-
Planck-Gesellschaft, and members of the scientic community. As a result these representatives agreed to support the
proposal in order to I) demonstrate that also a large-scale research facility can be realized at a university, II) withstand
a further emigration of such research from the universities, and III) nd a constructive solution that could serve as
a model for university research. As a matter of fact such a solution was found in the following years. However, it has
to be said that the full nancial support would never have arrived if the RTM had not been designed stage by stage,
and each time delivered in perfect shape (often to the surprise of outside experts) by Helmut Herminghaus and his
crew of physicists and technicians. The rst stage of the RTM (14 MeV) went into operation already in May 1979, the
second stage (183 MeV) followed in 1983, and the last stage was ready for the experiments in the fall of 1990.
At present the microtron delivers a continuous beam of an intensity of about 100 A for unpolarized and 40 A for
polarized electrons with a polarization degree of about 80 %. Its energy close to 1 GeV provides the perfect resolution
to study the distributions of charge, magnetization, and strangeness inside the nucleon and light nuclei, the threshold
production of the Goldstone bosons pion and eta, the polarizabilities of nucleons and pions, and the excitation of
the most prominent nucleon resonance, the (1232). Since the physics with the rst two stages of the RTM was
summarized already at an earlier workshop ( Physics with MAMI A ), the present Symposium concentrates on the
achievements of the years with the 855 MeV stage (MAMI B). The organizers also decided to invite as speakers, with
a few exceptions, young colleagues who have made a career with their work at MAMI.
It remains to say thank you to many people and institutions for continuing support. We are grateful to all the
colleagues from the Institute, the postdocs, Ph.D. and younger students who contributed to the MAMI project.
VIII
Special thanks go to the people in the workshops and in the administration without whose eorts the project could
never have succeeded. We are grateful to the colleagues from the neighbouring Institut fur Physik for their work on
polarized beams and targets, for the TAPS detector brought to Mainz by the Gie en group, to the Bonn/Bochum
group for the polarized H2 -target, and to many other German institutions for active engagement and various detection
devices, notably Darmstadt, Erlangen, Gottingen, and Tubingen. Our thanks go to the foreign colleagues who have
participated in the project from the very beginning, notably to our Scottish colleagues who built the photon tagger
with the support of their SERC, the groups from Pavia sponsored by the INFN, from Saclay supported by the
CEA/DAPNIA and from Orsay supported by the CNRS. We appreciate common experimental and theoretical work
with physicists from various other places in Europe, e.g. Amsterdam (NIKHEF), Basel, Genova, Gent, Lljubljana,
Trento and several Russian universities and institutions, and from overseas, e.g., Jeerson Lab, MIT, Florida State
University, University of Nagoya, George Washington University, and TRIUMF. Finally, in view of the upgrade two
more collaborations have developed in recent years. The Crystal Ball Collaboration has shipped its detector from the
Brookhaven National Lab to Mainz, and the KAOS detector is being installed in Mainz with the help of the GSI
Darmstadt. Last but not least we are grateful to the members of the international Program Advisory Committee and
of numerous evaluation and expert committees for their invaluable scientic advice and moral uphold.
Concerning the institutions we rst and foremost thank our Physics Faculty, the Johannes Gutenberg-Universitat
and the State of Rhineland-Palatinate for continued and coherent support. We are extremely grateful to the state
and to the federal ministries (BMFT, BMBW, BMBF) who nanced the construction of the new accelerator and
experimental halls as well as the large spectrometers via the university construction program (HBFG). Our special
thanks go to the Deutsche Forschungsgemeinschaft that backed up the project by means of Collaborative Research
Centers (SFB 201, CRC 443) whose resources were of the utmost importance to sustain our postdoc and PhD program.
Finally, we received recent support by the European networking activities via the I3HP/Transnational Access program.
Last but not least the organizers are grateful to the speakers of this Symposium for summarizing the various
achievements with MAMI and related research, and for bringing back memories of the past. Though retirees enjoy
the latter aspects very much, there is no reason to engage in retrospection: The double-sided microtron is expected to
yield its 1.5 GeV electron beam later this year, and we wish our colleagues and their students all the success in the
years to come!
Hartmuth Arenhovel
Hartmut Backe
Dieter Drechsel
Jorg Friedrich
Karl-Heinz Kaiser
Thomas Walcher
The Editors
Eur. Phys. J. A 28, s01, 1 5 (2006)
DOI: 10.1140/epja/i2006-09-001-x EPJ A direct
electronic only
/
Published online: 15 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Precision experiments using the electromagnetic probe have recently produced important new
data on fundamental properties of the nucleon, e.g. charge, magnetism, shape, polarizability, spin and
sea quark structure. These experiments have been made possible by a new generation of high duty factor
electron accelerators, advances in spin polarization technology (beams, targets and recoil polarimeters),
and the development of unique, optimized detector systems. In this contribution, the role of multiple
photon exchange in electron scattering from the proton and the role of sea quarks in nucleon structure are
highlighted.
PACS. 13.40.Gp Electromagnetic form factors 13.60.-r Photon and charged-lepton interactions with
hadrons 13.60.Fz Elastic and Compton scattering 14.20.Dh Protons and neutrons
2 Evidence for multiple photon eects in of multiple photon exchange and so give an incorrect de-
elastic electron scattering from the proton termination at higher Q2 , i.e. above about 1 (GeV/c)2 .
This multiple photon exchange contribution to elas-
tic electron-proton scattering can be conrmed by precise
Essentially all electron scattering experiments to study
comparison of electron-proton with positron proton elas-
proton and nuclear structure to date have been analyzed
tic scattering or by measurement of the asymmetry Ay in
in terms of single photon exchange. The ne structure cou-
scattering of unpolarized electrons from a vertically polar-
pling constant 1/137 is small enough that leading or-
ized proton target [21]. If conrmed, this is a very signi-
der has been adequate. There are a few specic examples
cant result.
where multiple photon exchange is known to be signif-
icant, e.g. in comparison of electron and positron scat-
tering in kinematics where the single photon exchange
cross-section is small, or in radiative processes. Thus, it 3 Role of sea quarks in nucleon structure
came as a surprise when the Jeerson Lab Hall A recoil
polarization measurements of electron-proton elastic scat- QCD tells us that the nucleon comprises three valence
tering at momentum transfers of about 2 (GeV/c)2 [18] quarks and a sea of quark-antiquark pairs. From the ear-
showed a substantial deviation from the data obtained liest days of nuclear physics, these sea quarks in the form
over several decades with the Rosenbluth technique [19], of mesons, have been viewed as playing an important role
which is based on precise cross-section measurements. in the long distance structure of the nucleon e.g. the mag-
This discrepancy has been interpreted as the eect of mul- nitude and sign of the proton and neutron magnetic mo-
tiple photon exchange in the elastic electron-proton cross- ments. In addition, the most successful hadronic theoreti-
section [20]. The cross section for elastic electron-proton cal descriptions of light nuclei incorporate meson exchange
scattering in the one-photon exchange approximation can between nucleons as an essential element of nuclear bind-
be written in terms of the pointlike Mott cross-section, the ing. This meson cloud structure to the nucleon has gen-
Sachs form factors GpE and GpM and the electron scattering erally been accepted but has lacked both a rigorous the-
angle as oretical underpinning and a denitive quantitative basis
from experiment.
The role of valence quarks in nucleon structure has
d d Gp2 + G p2
p2
= E M
+ 2 GM tan 2
, been studied extensively. The eects of sea quarks and
d d M ott 1+ 2 gluons are relatively poorly determined, in large part be-
cause they require high center-of-mass energy, and are a
where = Q2 /4M 2 . Figure 1 shows the recoil polarization major focus of interest for the future [22]. One of the im-
determination of GpE /GpM (solid circles) as a function of portant contributions over the last decade has been the
momentum transfer Q2 . The Rosenbluth data (all other experimental measurement of deep inelastic scattering at
data points) are believed to be uncorrected for the eects high energies to determine the eects of the sea quarks and
R.G. Milner: The beauty of the electromagnetic probe 3
5 Conclusion
The author would like to acknowledge discussions with A.M. 14. E.S. Ageev et al., Phys. Lett. B 633, 25 (2006).
Bernstein, T.W. Donnelly, R. Miskamen, A.H. Mueller, J.W. 15. J. Kiryluk (MIT) for the STAR collaboration, Proceed-
Negele, and C.N. Papanicolas. In addition, the author would ings of PANIC 2005, October 2005, Santa Fe, New Mex-
like to acknowledge that the BLAST experiment is the fruit of ico; K. Boyle (Stony Brook) for the PHENIX collabora-
a dedicated collaboration over an extended period of time. In tion, Proceedings of PANIC 2005, October 2005, Santa Fe,
particular, an outstanding cohort of graduate students is play- New Mexico, to be published by the American Institute of
ing an essential role. The authors research is supported by the Physics.
United States Department of Energy under the Cooperative 16. S. Kowalski, these proceedings.
Agreement DE-FG02-94ER40818. 17. F. Maas, these proceedings.
18. V. Punjabi et al., Phys. Rev. C 71, 055202 (2005).
19. I.A. Qattan et al., Phys. Rev. Lett. 94 (142301) (2005).
References 20. P.A.M. Guichon, M. Vanderhaeghen, Phys. Rev. Lett. 91,
142303 (2003); P.G. Blunden, W. Melnitchouk, J.A. Tjon,
1. R.G. Edwards et al., Phys. Rev. Lett. 96, 052001 (2006). Phys. Rev. Lett. 91, 142304 (2003).
2. D. Rohe et al., Phys. Rev. Lett. 83, 4257 (1999). 21. A.A. Afanasev et al., Phys. Rev. D 72, 013008 (2005).
3. T. Eden et al., Phys. Rev. C 50, R1749 (1994). 22. A. Deshpande, R. Milner, R. Venugopalan, W. Vogelsang,
4. I. Passchier et al., Phys. Rev. Lett. 82, 4988 (1999). Annu. Rev. Nucl. Part. Sci. 55, 165 (2005).
5. R. Madey et al., Phys. Rev. Lett. 92, 122002 (2003). 23. S. Chekanov et al., Phys. Rev. D 67, 012007 (2002).
6. N. Sparveris et al., Phys. Rev. Lett. 94, 122003 (2005). 24. J. Friedrich, Th. Walcher, Eur. Phys. J. A 17, 607 (2003).
7. V. Pascalutsa, M. Vanderhaeghen, Phys. Rev. Lett. 95, 25. BLAST Technical Design Report August 10th, 1997.
232001 (2005). 26. D. Cheever et al., Nucl. Instrum. Methods A 556, 410
8. C. Alexandrou et al., Phys. Rev. Lett. 94, 021601 (2005). (2006).
9. J. Roche et al., Phys. Rev. Lett. 85, 708 (2000). 27. A. Maschinot, MIT PhD Thesis 2005 (unpublished).
10. G. Laveissiere et al., Phys. Rev. Lett. 93, 122001 (2004). 28. H. Arenhovel, W. Leidemann, E.L. Tomusiak, Phys. Rev.
11. P. Bourgeois et al., submitted to Phys. Rev. Lett. (April C 52, 1232 (1995); 46, 455 (1992); Z. Phys. A 331, 123
2006). (1988); 334, 363 (1989).
12. A. Airapetian et al., Phys. Rev. D 71, 012003 (2005). 29. The G0 Collaboration, Phys. Rev. Lett. 95, 092001 (2005).
13. X. Zheng et al., Phys. Rev. C 70, 065207 (2004).
Eur. Phys. J. A 28, s01, 7 17 (2006)
DOI: 10.1140/epja/i2006-09-002-9 EPJ A direct
electronic only
/
Published online: 31 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. The Continuous Electron Accelerator Facility, CEBAF, located at the Thomas Jeerson Na-
tional Accelerator Facility, is devoted to the investigation of the electromagnetic structure of mesons,
nucleons, and nuclei using high energy, high duty-cycle electron and photon beams. Selected experimental
results of particular interest to the MAMI community are presented.
PACS. 29.17.+w Electrostatic, collective, and linear accelerators 25.20.-x Photonuclear reactions
25.30.Bf Elastic electron scattering 25.30.Dh Inelastic electron scattering to specic states
1 Personal Comments Dieter (Drechsel) has always been one of those peo-
ple I have looked to as the source of the big picture in
It is an honor and a pleasure to be here to celebrate the nuclear physics. He has provided us with deep insights, a
achievements of MAMI and the distinguished careers of sense of direction, and an understanding of what is really
Professors Arenhovel, Backe, Drechsel, Friedrich, Kaiser, important. He has also been an inspiring example here at
and Walcher. Mainz of the tremendous benets to everyone of having a
close collaboration between theory and experiment.
We are all deeply aware of the extent to which the
science we do builds on the achievements of those who Jorg (Friedrich) has taught us all how to analyze and
have gone before us, and on the insights and hard work interpret electron scattering data with minimal prejudice
of our colleagues working in the eld today. One of my (and, therefore, maximal honesty). It is a delight to see
very earliest memories as a scientist, dating from the days the same rigorous approach that was so successful in the
when I was a young graduate student, is that of attending study of nuclei and their excited states now being applied
Photonuclear Physics Boot Camp (otherwise known as to nucleon structure.
the Photonuclear Gordon Conference) and learning from Karl-Heinz (Kaiser) and his mentor, Helmut Herming-
(and with) many of those retiring today. haus, taught the world how to build superb continuous-
wave (cw) electron accelerators eectively and e ciently.
Thomas (Walcher) was one of the very rst scientists
Karl-Heinz, in particular, through the design and con-
I ever knew beyond the boundaries of my own laboratory.
struction of the double-sided microtron, is leaving the In-
He came to visit us (at Yale), and I went and visited him
stitute well positioned for another generation of superb
and his colleagues at Darmstadt. It has been a great plea-
experiments.
sure to follow his distinguished career in science, from low-
In conclusion, on behalf of so many people I have
Q2 electron scattering to hadronic physics at CERN and
worked with in nuclear physics, I want to thank each of you
beyond, and nally to the leadership role he has played at
for your many contributions to our eld, and to express
MAMI for many years.
the hope we all share that for each of you retirement is
Hartmuth (Arenhovel) has been the keeper of the ame a formality, not a reality, and that you will continue to be
of all knowledge about the deuteron, and a worthy succes- active for years to come.
sor to Gregory Breit. You should know that I was a grad-
uate student at Yale, and it was one of Professor Breits
missions in life to convince any and all who would listen
that the deuteron was the essence of nuclear physics, and 2 Research at Jeerson Laboratory
that until we understood the deuteron, we did not un-
derstand anything. I think it is fair to call Hartmuth the The Thomas Jeerson National Accelerator Facility, also
Gregory Breit of my generation; he has made so many called Jeerson Lab (or JLab), operates the Continu-
contributions. ous Electron Beam Accelerator Facility (CEBAF). CE-
BAF is a cw electron accelerator capable of delivering
a
e-mail: cardman@jlab.org three electron beams for simultaneous experiments in the
8 The European Physical Journal A
1.5
three experimental areas. Originally designed for 4 GeV,
GE
p
2
NIKHEF: H(e,en)
0.8 1.0
tent, SLAC), and by the enhanced research opportunities 0.8
provided by cw electron beams as demonstrated at fa- 0.6
0.6
cilities such as MAMI. CEBAF is a large, international 0.4 0.4
laboratory with a broad research program; it has been in 0 10
2
20
2
30 0 1 2 3 4
Q (GeV ) Q2 (GeV2)
operation for some seven and a half year now.
What are the goals of CEBAFs research program? Ba- Fig. 1. Nucleon form factor data available before the start of
sically, we aim to understand strongly-interacting matter. experiments using CEBAF. Top 2 panels: electric form factors,
How are the hadrons constructed from the quarks and glu- bottom panels: magnetic form factors. Left 2 panels: proton
ons of QCD, and how does the nucleon-nucleon force arise form factors, right panels: neutron form factors (adapted from
from the strong interaction? We further aim to identify ref. [1]).
the limits of our understanding of nuclear structure by us-
ing the high precision attainable with the electromagnetic
probe and the possibiltiy of extending investigations to What are the spatial distributions of the u, d, and s
very small distance scales. A specic issue that motivated quarks in the hadrons?
the construction of CEBAF was our desire to gain insight What is the excited state spectrum of the hadrons,
into the question of where the description of nuclei based and what does it reveal about the underlying degrees-
on nucleon and meson degrees-of-freedom fails and the un- of-freedom?
derlying quark degrees-of-freedom must be taken into ac- What is the QCD basis for the spin structure of the
count. One can ultimately characterize all of this as trying hadrons?
to understand QCD, not in the perturbative regime acces-
sible at very high energies and very short distance scales,
but in the strong interaction regime relevant to most of 3.1 What are the spatial distributions of the u, d, and
the visible matter in the Universe. To make progress in s quarks in the hadrons?
these areas, there are other critical issues that must be
addressed, such as the mechanism of connement, the dy- Elastic electron scattering has provided most of our infor-
namics of the quark interaction, and how chiral symmetry mation on the spatial distributions of the quarks in the
breaking occurs. nucleons. The data on the four electromagnetic structure
To provides some shape and structure to the discussion functions of the nucleon, GE and GM for both the proton
of the experiments, the CEBAF program can be organized and the neutron, available just prior to the start of exper-
into half a dozen broad thrusts. This presentation will iments at CEBAF is shown in g. 1. The magnetic form
concentrate on two of them: factors of the proton and the neutron were known rea-
sonably well, but the electric form factors were not. The
How are the nucleons made from quarks and glue? electric form factor of the proton had not been determined
Where are the limits of our understanding of nuclear accurately enough to distinguish between a wide range of
structure theories based on rather dierent physics. First results on
the electric form factor of the neutron were available from
Bates, Mainz, and NIKHEF, but these data were limited
3 How are the nucleons made from quarks to moderate momentum transfers and, therefore, not sen-
and glue? sitive to the details of the distribution of charge inside the
neutron. The measured form factor was consistent with
Among the most interesting puzzles in physics today are: the r.m.s. radius derived from neutron-electron scattering.
why there is this eective degree-of-freedom in QCD, the The present status of the nucleon form factors is
nucleon; and how something as complicated as the resid- shown in g. 2. The measurements of the polarization
ual QCD interaction between quarks in nucleons can be transfer from the incident electron to the elastically
characterized by a rather simple N-N potential? To pro- recoiling proton have shown that the electric and mag-
vide experimental insights that will help us solve the rst netic form factors for the proton dier substantially. The
of these puzzles, the Jeerson Lab research community has systematic dierences between the polarization transfer
mounted an array of investigations in three broad areas: data and the Rosenbluth results for GE /GM are likely
L.S. Cardman: Physics at the Thomas Jeerson National Accelerator Facility 9
1.5
GE
p
GMp 1.0
0.10 JLab E93-038:
2
MIT-Bates: H(e,en)
2
H(e,en) JLab E93-026: 2H(e,en)
3
Mainz A1: He(e,en)
n 0.08 Mainz A3: 2H(e,en)
GE 3
Mainz A3: He(e,en)
2
NIKHEF: H(e,en)
0.06 Schiavilla & Sick:GQ
0.5
0.04
0.02
Galster New Fit
0.0 0
0 2 4 6 0 1 2
2
Q (GeV ) 2 Q2(GeV2)
1.2 Borkowski
p
GM Sill
Bosted GM
n this measurement
Lung
Rock
(with statistical errors) Bartel
Walker
1.4
pGD 1.0 Andivahis
n G D
Arnold
Jourdan1
Jourdan2
1.2 Gao
Xu
0.8 1.0
0.8
0.6 grey: estimated systematic
0.6 error
0.4 0.4
0 10 20 30 0 1 2 3 4
2 2
Q (GeV ) Q (GeV2)
2
Fig. 2. Present status of the nucleon form factor data including the CEBAF data (adapted from ref. [1]).
1.5 0.10
due to two-photon exchange eects modifying the results. pGE
p
n
Theoretical estimates suggest that the modications are GMp
GE
1.0
much smaller for the polarization transfer data than 0.05
logical way with minimum prejudice [2], what emerges is Hammer (VMD + disp. rel.)
some of the clearest evidence we have for the nucleons Fig. 3. Theoretical descriptions of the nucleon form factor
pion cloud (see g. 4). Similar results have been obtained data [1].
using a dierent approach to model-independent analy-
sis [3] of nucleon form factors.
We plan to extend the proton form factor data to
9 (GeV/c)2 , where we may see evidence for a diraction historically proven to be important, and we expect these
minimum. The neutron form factor will also be extended data will provide further insight and sensitivity for com-
to 5 (GeV/c)2 . Further extensions of a factor of two are pleting our understanding of how to construct nucleons
planned with the 12 GeV Upgrade. Such extensions have from quarks and gluons.
10 The European Physical Journal A
0.03
r 2 ( r; GEn ) / fm-1
bn p0
0.02 (1- bn)n0 G0
total HAPPEx
0.01
A4
0.00
-0.01
-0.02
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
r / fm
0.03
+ Fig. 5. Nucleon strange form factor data from parity-violating
r ( r; GEppol ) / fm-1
0.02
-bpp0
bpn0
electron scattering at forward angles as a function of Q2 .
total
0.01
G0
(ex A4 SAMPLE with
0.00
tra GA calculation
po
lat
ed
-0.01
0.1 HA )
PP
2
E
GsE
X-
H
-0.02 ('0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4) 3
95% C.L.
r / fm HAPPEX-4He ('04) 2
0 1
+
Fig. 4. Neutron radial charge distribution (top) and the polar-
ization term contribution to the proton radial charge distribu-
tion (bottom) as inferred from an analysis using phenomeno-
logical models [2] of the world nucleon form factor data. The
-0.1
green solid lines show the pion cloud contribution. Note that Leinweber
r2 (r) is plotted, emphasizing the distribution at large radii. et al. (2005)
Q2 = 0.1 GeV2
GsM
-1 0 1
The strange quark form factors have become an in-
teresting area of study that is both analogous and com-
plementary to the classical electromagnetic form factors. Fig. 6. Separated strange nucleon form factors at Q2 =
By using the weak component of the electro-weak interac- 0.1 (GeV/c)2 [10].
tion we access the weak neutral current form factor, which
can be interpreted very elegantly in terms of the strange
quark distribution. Because there are no valence strange separating them spatially. Even at the highest momentum
quarks, this measurement provides a unique window on transfers reached experimentally we are averaging over a
the sea quark distribution. The strange form factors can distance scale that is roughly the size of the nucleon, so it
also be expected to provide us with interesting experi- is not too surprising that the result is small. There is an
mental insights into nucleon structure: by combining the intriguing suggestion in the data for something that one
electromagnetic and the weak neutral current form factors would call vaguely pion-cloud like behavior, but it is fair
we should be able to separate the spatial distribution of to say that the statistical signicance of this eect is not
the u, d, and s quarks. very high.
Figure 5 shows the worlds data on the strange proton The data taken at forward angles includes a mixture of
form factor taken at forward angles as a function of Q2 . electric and magnetic form factors. At Q2 = 0.1 (GeV/c)2
One sees data from the A4 experiment at Mainz [6,7,8], we have data at both forward and backward scattering
from HAPPEx I and II (JLab Hall A) [9,10,11], and from angles, so we can separate these eects (see g. 6). The
G0 (JLab Hall C) [13]. These di cult experiments would data favor a positive value of GsM , which is at variance
be impossible without highly polarized electron beams with most of the theoretical models. Experiments are in
from magnicently stable accelerators. The fact that the progress that will reduce the size of the error ellipse at
data from dierent laboratories lie roughly on a smooth this Q2 value by a about a factor of 3, and additional
curve gives one condence that the experimenters are do- experiments planned at both MAMI and JLab will permit
ing it right. separations at other Q2 values. A broad, world-wide eort
The rst thing that strikes you about the data is that will provide the results we want.
the form factor is rather small. This is to be expected, as Another interesting experiment is the measurement of
all of the strange quarks emerge as quark-antiquark pairs the pion form factor. The pion is the simplest QCD bound
popping in and out of the vacuum, and to get a nite system, the positronium of QCD. One expects that the
form factor there must be some kind of a polarizing eect pion form factor will provide us with evidence for the
L.S. Cardman: Physics at the Thomas Jeerson National Accelerator Facility 11
0.6
Amendolia +e elastics
+
Previous p(e,e )n
Q F
BSE+DSE
0.4 QCD Sum Rule
IMIN ARYCQM
0.2
PREL
Perturbative QCD
Missing States
W(GeV)
2 thresh.
N(1680)
N(1520)
(1232)
Mx(GeV)
80
)
A 1/2 (10 -3GeV -1/2)
-1/2
150
60
-3
50 20
0 0
-50 -20
-40
-100
P11(1440) -60
P11(1440)
-150
0 1 2 3 4 0 1 2 3 4
Q2 (GeV2) Q2 (GeV2)
Fig. 12. A 1 (Q2 ) and S 1 (Q2 ) multipoles for the P11 (1440)
2 2
Roper resonance [18].
2 2 1.0 1.2
-1. < Cos(K) < 0., Q = 0.7 (GeV/c)
at al.)
JLab E99-117 (3He)
C
al. HH
0.15 HF perturbed QM
T + L L E142 (3He)
pQCD
et ith
mas
World Data parm Q2 = 10 GeV2
)
de fit w
0.8 E154 (3He) 1.0
(Tho
Symmetric Q Wave function
n p
(L CD
r
Helicity 3/2 suppression
A1 HERMES (3He) A1
pQ
ea
model
Spin 3/2 suppression
)
al.
r)
gu
et
Bag
0.6
(Is
el
0.8
eig
M
CQ
(W
C
HH
n
to
o
w/ l.)
oli
lS
fit et a
0.1 0.4 SU(6)
ira
CD der 0.6
Ch
Q
p ea
(L
b/sr
Statistical model
0.2 (Soffer et al.)
0.4
CLAS-EB1b Q2 = 1.4 - 4.52 GeV2
HERMES
0 SMC
0.2
0.05 SLAC - E155
SLAC - E143
-0.2
0
known New
-0.4
N* N* ? 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x x
1.6 1.7 1.8 1.9 2 2.1
W (GeV) Fig. 14. (Left) Spin structure function of the neutron, An 1,
derived from 3 He data [19]; and (Right) spin structure function
Fig. 13. W-dependence of the cross section for the p(e, e K + ) of the proton Ap1 [20].
reaction integrated over backward-going K + [15].
s,g1 / JLab
0.16 Burkert-Ioffe
SLAC E143 2 s,g1 / world data GDH constraint
CLAS EG1a
1 p (no elastic)
(Q)/
0.08
s
0.06
0.04 0.1
0.02
0.05 -1
10 1 10
0
Q (GeV)
-0.02 Fig. 18. Eective strong-coupling constant as a function of Q2
0 0.5 1 1.5 2 2.5 3 derived from the Q2 -dependence of the Bjorken integral [23].
Q (GeV )
2 2
Fig. 16. Integral of the spin structure function of the proton of very low momentum transfer with high precision to test
as a function of Q2 [22]. the predictions of PT.
The rst signicant measurement of the Q2 -depen-
dence of the Bjorken integral (see g. 17) was made for
Q2 = (0.05 2.5) (GeV/c)2 . Remarkably, pQCD-based
0.20
Q2 evolution matches the data down to a Q2 of about
0.7 (GeV/c)2 . Deur et al. [23] have made an interesting in-
p-n
1
3 a) 1e-09
r(4He/3He)
Experiment
Correlated Basis Function
S(Em,pm) [MeV sr ]
Theory
-1
2 1e-10
-4
1 1e-11
4 0.250
pm
b)
r(12C/3He)
(G
3
eV
1e-12 0.410
c)/
2
0.570
1 1e-13
6 c)
0 0.1 0.2 0.3
Em (GeV)
r(56Fe/3He)
1 1.5 2 2.5
xB
100 10-1
previous data previous data
Conventional
10-1 Hall A MMD, S2, D
10-2 Nuclear Theory
Hall C MMD, 0, D
MMD, S2, D Forest and Schiavilla, IA
10-2
MMD, 0, D 10-3 Hall A
Forest and Schiavilla, IA
10-3 Forest and Schiavilla, IA+pair
10-4
10-4
2
A(Q ) 10-5
10-5
10-6
B(Q2)
10-6
10-7
10-7
10-8 10-8
10-9 10-9
0 2 4 6 0.0 0.5 1.0 1.5 2.0 2.5
Q2 (GeV 2 ) Q2 (GeV 2 )
1.0
0.5
t20(Q2)
0.0
VEPP 1985
Bates 1984
-0.5 VEPP 1990
Bates 1991
NIKHEF 1995
NIKHEF 1996
-1.0 JLab/POLDER
MMD, S2, D
MMD, 0, D
9. K.A. Aniol et al., Phys. Rev. Lett. 96, 022003 (2006). 20. CLAS Collaboration (K.V. Dharmawardane et al.), sub-
10. K.A. Aniol et al., Phys. Lett. B 635, 275 (2006). mitted to Phys. Rev. Lett. (2006).
11. K.A. Aniol et al., Phys. Lett. B 509, 211 (2001). 21. A. Thomas, these proceedings.
12. K.A. Aniol et al., Phys. Rev. C 69, 065501 (2004). 22. A. Deur, 13th Int. workshop on Deep Inelastic Scattering
13. D.S. Armstrong et al., Phys. Rev. Lett. 95, 092001 (2005). (DIS2005), AIP Conf. Proc. 792, 969 (2005).
14. T. Horn, private communication for the F collaboration. 23. A. Deur, V. Burkert, J.P. Chen, W. Korsch, arXiv:hep-
15. V.D. Burkert, Eur. Phys. J. A 17, 303 (2003). ph/0509113.
16. L.C. Smith, Invited Talk, Japan-US Workshop on Electro- 24. K. Egiyan et al., Phys. Rev. Lett. 96, 082501 (2006).
magnetic Meson Production and Chiral Dynamics, Osaka, 25. D. Rohe et al., Phys. Rev. Lett. 93, 182501 (2004).
Japan (April 2005). 26. F. Benmokhtar et al., Phys. Rev. Lett. 94, 082305 (2005).
17. CLAS Collaboration (V.D. Burkert), Int. J. Mod. Phys A 27. M. Garcon, J.W. Van Orden, Adv. Nucl. Phys. 26, 293
20, 1531 (2005). (2001) and references therein.
18. I. Aznauryan talk at N*2005, Tallahasse, FL (October 28. M. Mirazita et al., Phys. Rev. C 70, 014005 (2004).
2005), to be published. 29. E.L. Schulte et al., Phys. Rev. Lett. 87, 102302 (2001).
19. X. Zheng et al., Phy. Rev. Lett. 92, 012004 (2004).
Eur. Phys. J. A 28, s01, 19 27 (2006)
DOI: 10.1140/epja/i2006-09-003-8 EPJ A direct
electronic only
/
Published online: 10 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Few-body systems provide a testing ground for models of the NN interaction, reaction mecha-
nisms and for models of nuclei. An overview of results of coincidence experiments on the deuteron, 3 He
and 4 He obtained in the last 20 years at MAMI and at other facilities, covering a wide range of momentum
and energy transfers, is presented.
PACS. 25.10.+s Nuclear reactions involving few nucleon systems 25.30.-c Lepton-induced reactions
25.30.Fj Inelastic electron scattering to continuum
pf 10
-4
-5
e 10
exp
2
e 10
-8
pm 10
-9
5
-10
10
Fig. 2. Geometry and kinematical variables for the (e, e N)
reaction. 10
-11
-12
shown in g. 1. In the case of the Plane-Wave Impulse Ap- 10 0 200 400 600 800 1000
proximation (PWIA) the virtual photon is absorbed by a pm [MeV/c]
bound nucleon having a certain initial momentum pi . The
struck proton subsequently leaves the nucleus with a nal Fig. 3. D(e, e p)n cross section measured at MAMI [11] com-
momentum pf . The residual system may remain in its pared to calculations by H. Arenhovel [13].
ground or in an excited state and has a recoil momentum
pm . In the following, the term missing momentum will be
used synonymously with recoil momentum. Within PWIA tion of the (e, e N) cross section into an elementary (o-
the following relation between initial and missing momen- shell) electron nucleon cross section [8] and the spectral
tum is valid: pm = pi . The transferred energy is di- function describing the probability of nding a nucleon
vided between the kinetic energy of the ejected nucleon, its with a given initial momentum and missing energy. In-
separation energy, and the kinetic and, possibly, excitation tegrating the spectral function over the missing energy
energy of the residual system. The missing momentum pm leads to the momentum distribution. Final state interac-
and missing energy Em are dened as follows: tions (FSI), MEC, and IC remove this simple relation and
therefore pi = pm (g. 1).
Momentum conservation : q = pf + pm ,
Energy conservation : E m = T p Tr .
Here Tp is the kinetic energy of the ejected nucleon, and 3 Studies of the deuteron
Tr is the kinetic energy of the recoiling system, calculated
from pm under the assumption that the undetected (A- Early (e, e p) experiments were limited in luminosity by
1)-system remains in its ground state. the duty factor of the available electron accelerators. Cross
Figure 2 shows the electron scattering plane, dened sections could be measured for large missing momenta
by the incoming and scattered electron momenta, and the (pm 0.5 (GeV/c)) only at relatively small momentum
reaction plane, dened by the nal nucleon momentum transfer (Q2 0.1 (GeV/c)2 ) or for large Q2 only at rel-
and the momentum transfer. The cross section in the one atively small (pm < 0.2 (GeV/c)).
photon exchange limit can be written as [5,6,7] Experiments have been carried out at all facilities men-
tioned above. More recent experiments, carried out in the
d5 last ten years, beneted from the availability of high duty
= M ott (vL RL + vT RT + cycle beams at Jeerson Lab, MAMI, NIKHEF(AmPS),
dde dp
and at MIT-Bates(SHR). In general, the various exper-
+ vLT RLT cos + vT T RT T cos 2),
iments can be separated into those that explored the
where Ri are the response functions containing matrix ele- D(e, e p)n cross section over a large range of missing mo-
ments of the charge and current operators. These, in turn, menta and those that extracted individual response func-
provide the nuclear structure information. The vi are kine- tions.
matical factors depending on the electron kinematics only,
and M ott is the Mott cross section describing the scat-
tering of relativistic electrons by a point charge. 3.1 Cross section measurements at low Q2
If one neglects the interaction of the outgoing nucleon
with the recoiling system, one obtains the plane-wave im- These experiments explored the D(e, e p)n cross section
pulse approximation (PWIA) which permits a factoriza- over a wide range of missing momenta at small to medium
W.U. Boeglin: Few-nucleon systems at MAMI and beyond 21
120 8
Saclay 7
110 Bates NIKHEF
f00exp/Aren (%)
6 Bates
100 5
f00(fm)
1.2 4
90 3
2
80
exp/NMIR
1
1 70 0
-100 -50 0 50 100 20 30 40 50 60 70
120 8
Saclay 7
110 Bates NIKHEF
f11exp/Aren (%)
0.8 6 Bates
100 5
f11(fm)
4
90 3
0.6 2
80
1
70 0
0 50 100 150 200 250 300 -100 -50 0 50 100 20 30 40 50 60 70
pmiss (MeV/c) pm (MeV/c) pm (MeV/c)
Fig. 4. Ratio of experiment to theory for low missing mo- Fig. 5. Comparison of RL (f00 ),RT (f11 ) measurements to
menta. Data points: blue diamonds [11], red squares [14], vi- one another and theory. Left: Ratio (in %) of the experimental
olet circles [12], turquoise triangles down [9], green triangles response functions from Saclay [14] and Bates [15] to the cal-
left [15], orange triangles up [16]. culation by H. Arenhovel [19], [20]. Right: Response functions
from NIKHEF [17, 18] and Bates [15].
PWBA
N and has the additional advantage that the absolute cross
N + MEC + IC
10 N + MEC + IC + REL section normalization cancels in the ratio. An overview of
Data
experimental results is shown in g. 7 together with the
1 result of a determination of ALT at MAMI [23]. Other
recent ALT measurements have been published in refer-
RL / fm
0.01 D(e, e p)n cross sections have been obtained at SLAC for
high Q2 but low recoil momenta (pm < 0.2 GeV/c) [30].
0.001 Recently, experiments have been carried out at Jeerson
0 50 100 150 200 250 300 350 400
pm / (MeV/c) Lab in Hall A (experiment E01-020) as well as in Hall B
using CLAS (experiment E94-019). The goal of the Hall
Fig. 6. Result of a RL -RT separation at MAMI [21] for miss-
A experiment is to test the Generalized Eikonal Approx-
ing momenta up to 350 MeV/c at a momentum transfer of
imation (GEA) description of FSI [31] in the D(e, e p)n
450 MeV/c. Calculations are by H. Arenhovel [20] (PWBA: no
FSI but scattering o the neutron and observing the recoiling
reaction while the goal of the Hall B experiment is to
proton has been included). use the GEA description of the D(e, e p)n reaction in
the search for evidence of color transparency. Within the
GEA, FSI are described by a series of small-angle scatter-
0.4
exp
ings of the outgoing nucleon. This approximation, which
0.2 N+MEC+IC+REL
PWIA is typically valid for nucleon energies of 1 GeV and above,
0 has been successfully applied in high-energy nucleon scat-
tering. However it has never been tested for the D(e, e p)n
ALT
-0.2
A
-0.4
reaction. Another goal of the Hall A experiment is the
determination RLT for missing momenta up to 0.5 GeV/c
-0.6
where relativistic eects are expected to be very large and
-0.8
0 100 200
Pmiss (MeV/c)
300 RLT is sensitive to details of the current operator.
The GEA predicts a characteristic dependence of the
Fig. 7. Left: an overview of measurements of ALT from var- strength of FSI on the angle of the recoiling neutron with
ious experiments [25] where 0.15 Q2 0.22 (GeV/c)2 . respect to the momentum transfer and on the value of the
Calculations are from H. Arenhovel et al. [26] (dash-dot: missing momentum. For angles around 80 FSI eects are
N+MEC+IC, solid: N+MEC+IC+RC) and from E. Hummel predicted to be maximal. For pm = 0.2 GeV/c a reduction
et al. [27] (dashed: PWBA), the dotted curve corresponds to of the cross section by about 30 40% is predicted and for
PWIA (cc1 ). 0.15 to 0.22. Right: Determination of ALT at pm = 0.4 GeV/c and pm = 0.5 GeV/c an increase of the
MAMI for Q2 = 0.33 (GeV/c)2 [23]. The calculation is by cross section by more than a factor of two is predicted. The
H. Arenhovel [20] including FSI, MEC, IC and Relativistic location of the extremum of the rescattering contributions
corrections (RC). give additional information about the details of the rescat-
tering process such as the importance of the Fermi motion
of the bound nucleons.
angle is changed (in the electron scattering plane) in such In order to address these questions in Hall A, the
a way, that the reaction plane varies between = 0 and D(e, e p)n cross section has been measured for Q2 =
= 180 (see g. 2). The cross section dierence obtained 0.8, 2.1 and 3.5 (GeV/c)2 and missing momenta between
from these two measurements is then proportional to RLT . pm = 0 and pm = 0.5 GeV/c. A very preliminary result is
A quantity closely related to RLT is the left-right asym- shown in g. 8 where the observed yield is compared to a
metry Monte Carlo calculation using the PWIA for the cross sec-
0 tion. Clearly the predicted angular dependence has been
ALT = 180
180 + 0 observed. A detailed analysis is currently in progress.
W.U. Boeglin: Few-nucleon systems at MAMI and beyond 23
5
FSI / PWIA
4.5
4 pm=0.5 GeV/c exp.
3.5
3 pm=0.5 GeV/c + MEC
pm= 0.5 GeV/c no MEC
2.5
pm= 0.2 GeV/c exp.
2
1.5
1
0.5 pm= 0.2 GeV/c no MEC
0
0 20 40 60 80 100 120 140 160 180
nq
4 3 He and 4 He studies
The breakup of the 3,4 He nuclei can lead to a 2-body -
nal state like in the deuteron or it can lead to a 3- and
even 4-body nal state for 3 He and 4 He, respectively. Only
recently, with the advent of high computing power and ef-
cient computational techniques, can the continuum nal
state be calculated accurately. These nuclei are the sim- Fig. 10. The asymmetry ALT and the response function RLT
plest systems in which to study short range correlations. measured at MAMI for the 4 He(e, e p)3 H reaction [39] together
with calculations with and without the inclusion of MEC [40].
-6 0.75
10
PWIA
-7 0.50 GLB
10 PWIA
GLB+MEC
GLB
-8 GLB+MEC 0.25
10
(fm /MeV/sr )
2
-9
10 0.00
ALT
10
-10 -0.25
2
10
-11 -0.50
10
-12
-0.75
-13
10 -1.00
-14
10 -1.25
0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200
pm(MeV/c) pm(MeV/c)
d d p [ b/MeVsr2]
have been compared to calculations by R. Schiavilla et 10-4
al. [40,41], which show the need for MEC to improve the
agreement with the experiment (g. 10).
10-5
1
Tp1 (a) (b) Data
150 Cut Data
Phase Space
Counts
100
0.5
50
0 0
0 0.5 1 -1 -0.5 0 0.5 1
Tp2
cos( )
pp
Fig. 16. a) Dalitz plot for the lab frame Tp1 / versus Tp2 /
for events with pN > 0.25 GeV/c [49]. b) The cosine of the
p-p lab frame opening angle. Open circles for events with small Fig. 18. Polarization transfer experiments on 4 He from
proton momenta (Tp < 0.2 ) and closed circles show all data. Mainz [55] and JLAB [56]. The best agreement is obtained for a
modied nucleon form factor within the quark-meson coupling
e model [57, 58].
P y
e pf
P x 1.05
q P z
1.00
e
(P x /P z )/(Px /Pz )PWIA
0.95
Fig. 17. Denition of polarization variables for (e, e N) polar-
ization transfer experiments. e represents the polarized elec-
tron beam, e the scattered electron and e the electron scatter-
ing angle. q the virtual photon, pf the nal nucleon momentum 0.90
and Px , Py , Pz the polarization of the ejected nucleon.
4
He EXP
1
0.85 H EXP
OPT(no CH EX)
the relative angular distribution of the observed nucleon OPT
pair that shows a pronounced peak at 180 (in the Lab OPT+MEC
JLAB [56] on 4 He showed that the ratio 2. R. Schiavilla, V.R. Pandharipande, R.B. Wiringa, Nucl.
Phys. A 449, 219 (1986).
(Px /Pz )He 3. J. Carlson, Phys. Rev. C 36, 2026 (1987).
R=
(Px /Pz )H 4. B.S. Pudliner et al., Phys. Rev. Lett. 74, 4396 (1995).
5. A.E.L. Dieperink, T. de Forest, Annu. Rev. Nucl. Sci. 25,
is reduced by about 10% (g. 18). Some calculations based 1 (1975).
on the quark meson coupling model [57,58] suggest a pos- 6. S. Frullani, J. Mougey, Adv. Nucl. Phys. 14, 1 (1984).
sible modication of the nucleon form factors inside nuclei 7. S. Bo , C. Giusti, F.D. Pacati, Phys. Rep. 226, 1 (1993).
8. T. de Forest, Nucl. Phys. A 392, 232 (1983).
to account for this observation. Another calculation by 9. M. Bernheim et al., Nucl. Phys. A 365, 349 (1981).
Schiavilla et al. [59] uses realistic wave functions for the 10. S. Turck-Chieze et al., Phys. Lett. B 142, 145 (1984).
bound state that include correlation eects. One- and two- 11. K.I. Blomqvist et al., Phys. Lett. B 429, 33 (1998).
body currents are included and spin and isospin depen- 12. P.E. Ulmer et al., Phys. Rev. Lett. 89, 062301 1 (2002).
dences in the nal state interaction including charge ex- 13. H. Arenhovel, W. Leidemann, L. Tomusiak, Phys. Rev. C
change have been taken into account. The calculation can 52, 1232 (1995).
reproduce the experimental data without the need of form 14. J.E. Ducret et al., Phys. Rev. C 49, 1783 (1994).
factor modications as can be seen in g. 19. New high- 15. D. Jordan et al., Phys. Rev. Lett. 76, 1579 (1996).
16. Kasdorp et al., Few-Body Syst. 25, 115 (1997).
precision experiments on 4 He are planned at Jeerson Lab
17. M. van der Schaar et al., Phys. Rev. Lett. 66, 2855 (1991).
in Hall A to improve and extend the available data. 18. M. van der Schaar et al., Phys. Rev. Lett. 68, 776 (1992).
19. W. Fabian, H. Arenhovel, Nucl. Phys. A 314, 253 (1979).
20. H. Arenhovel, private communication (2001).
6 Summary and conclusion 21. R. Boehm, Thesis, University of Mainz (2001).
22. S. Jeschonnek, J.W. Van Orden, Phys. Rev. C 62, 044613
In the last 20 years much progress has been made in the (2000).
knowledge of the structure of few-body systems. The avail- 23. W.U. Boeglin, private communication (2005).
24. Z.-L. Zhou et al., Phys. Rev. Lett. 87, 172301 (2001).
ability of high quality CW beams made it possible to mea-
25. Z.-L. Zhou et al., Proceedings of the MIT-Bates Workshop
sure coincidence cross sections over a wide range of kine- (1998).
matical variables which were inaccessible before. In paral- 26. F. Ritz, H. Goller, Th. Wilbois, H. Arenhovel, Phys. Rev.
lel, theoretical progress together with increasing compu- C 52, 1232 (1995).
tational power has resulted in sophisticated models that 27. E. Hummel, J.A. Tjon, Phys. Rev. C 49, 21 (1994).
agree very well with the data. 28. A. Pellegrino et al., Phys. Rev. Lett. 78, 4011 (1997).
Coincidence data on 3 He have enjoyed a lot of atten- 29. S.M. Dolni et al., Phys. Rev. C 60, 064622 (1999).
tion and the analysis of recent data taken at Jeerson 30. H.J. Bulten et al., Phys. Rev. Lett. 74, 4775 (1995).
Lab is still in progress. The importance of a detailed un- 31. L.L. Frankfurt, M.M. Sargsian, M.I. Strikman, Phys. Rev.
derstanding of nal state interactions, MEC, and IC is C 56, 1124 (1997).
32. J.M. Laget, Phys. Lett. B 609, 49 (2005).
crucial in order to extract information on the short-range 33. C. Marchand et al., Phys. Rev. Lett. 60, 1703 (1988).
structure of light nuclei. 34. E. Jans et al., Nucl. Phys. A 475, 687 (1987).
While many new single-arm data on the deuteron have 35. R.E.J. Florizone et al., Phys. Rev. Lett. 83, 2308 (1999).
been obtained, available coincidence data and especially 36. A. Kozlov et al., Phys. Rev. Lett. 93, 132301 (2004).
the lack of new, high-precision response function determi- 37. J.M. Le Go et al., Phys. Rev. C 55, 1600 (1997).
nations are disappointing. 38. R.E.J. Florizone, Thesis, MIT (1999).
Triple coincidence experiments on the He nuclei are 39. K. Aniol et al., Eur. Phys. J. A 22, 449 (2004).
expected to provide new data on the structure of corre- 40. R. Schiavilla et al., Phys. Rev. Lett. 54, 835 (1990).
41. J. Forest et al., Phys. Rev. C 54, 646 (1996).
lations. However, these experiments are very complex to 42. M. Rvachev et al., Phys. Rev. Lett. 94, 192302 (2005).
carry out, analyze, and interpret. Several experiments are 43. R. Schiavilla et al., Phys. Rev. C 72, 064003 (2005).
being currently analyzed at MAMI and JLAB. The (up- 44. L.P. Kaptari, C. Cio degli Atti, Phys. Rev. C 71, 024005
graded ) Mainz Microtron will continue to play a leading (2005).
role in nuclear physics. 45. F. Benmokhtar et al., Phys. Rev. Lett. 94, 082305 (2005).
46. J.M. Laget, Few-Body Syst., Suppl. 15, 171 (2003).
47. D.L. Groep et al., Phys. Rev. Lett. 83, 5443 (1999).
I would like to thank H. Arenhovel, H. Backe, D. Drechsel, 48. D.L. Groep et al., Phys. Rev. C 63, 014005 (2000).
J. Friedrich, K-H. Kaiser and Th. Walcher for making MAMI 49. R.A. Niyazov et al., Phys. Rev. Lett. 92, 52303 (2004).
such a success, for their contributions to our eld and for giving 50. R. Arnold et al., Phys. Rev. C 23, 363 (1981).
me the opportunity to carry out research at MAMI. I wish you 51. M.K. Jones et al., Phys. Rev. Lett. 84, 1389 (2000).
all the best in the future. This work was supported in part by 52. O. Gayou et al., Phys. Rev. C 64, 038202 (2001).
the Department of Energy, DOE grant DE-FG02-99ER41065. 53. O. Gayou et al., Phys. Rev. Lett. 88, 092301 (2002).
54. T. Pospischil et al., Eur. Phys. J. A 12, 125 (2001).
55. S. Dietrich et al., Phys. Lett. B 500, 47 (2001).
56. S. Strauch et al., Phys. Rev. Lett. 91, 052301 (2003).
References 57. D.H. Lu et al., Phys. Lett. B 417, 217 (1998).
58. D.H. Lu et al., Phys. Rev. C 60, 068201 (1999).
1. A. Nogga, H. Kamada, W. Glockle, Nucl. Phys. A 689, 59. R. Schiavilla et al., Phys. Rev. Lett. 94, 072303 (2005).
357 (2001).
Eur. Phys. J. A 28, s01, 2938 (2006)
DOI: 10.1140/epja/i2006-09-004-7 EPJ A direct
electronic only
/
Published online: 11 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Experiments with polarized 3 He
at MAMI have already a long tradition. The A3 Collaboration
started in 1993 with the aim to measure the electric form factor of the neutron. At this time MAMI was
the second accelerator where experiments with 3 He
were possible. Some years before this pilot experiment
the development of the apparatus to polarize 3 He in Mainz started. There are two techniques which allow
for
to polarize sucient large quantities of 3 He. Both techniques will be compared and the benet of 3 He
at MAMI will be given
nuclear physics will be discussed. A review of the experiments done so far with 3 He
and the progress in the target development, the detector setup and the electron beam performance will be
pointed out.
PACS. 13.40.Gp Electromagnetic form factors 13.88.+e Polarization in interactions and scattering
25.70.Bc Elastic and quasielastic scattering 29.25.Pj Polarized and other targets
during this time in the development of the target, the vapor and large laser power (> 40 W) are needed to po-
performance of the polarized electron beam and the im- larize the gas in a target cell of 10 bar within 20 h to 50%.
provement of the detector setup will be demonstrated. The The advantage is that no further compression stage is re-
dierent objectives for the experiments will be presented. quired and a compact design is possible. To avoid radiation
Finally upcoming experiments at MAMI in the near future trapping in the optical thick Rb vapor which occurs when
are briey presented. unpolarized resonant uorescence light is emitted and af-
terwards reabsorbed, 50100 mbar nitrogen is added. The
addition of a fraction of 102 N2 leads to 5 (10)% con-
tribution to the scattering rate from a proton (neutron)
2 Polarization methods and therefore to an eective dilution of the polarization
observables.
For nuclear target applications two methods are in use, Except for experiments in a storage ring the mass
metastable-exchange optical pumping (MEOP) [15] and density of a few mbar of polarized 3 He from MEOP is
spin-exchange optical pumping (SEOP) [16]. Both meth- too low for a nuclear physics experiment. Therefore one
ods were already developed in the 60ies but became or two mechanical compression stages1 are necessary to
only ecient in use with the development of laser light reach pressures of up to 6 bars. Up to now three dierent
sources of sucient power and proper frequency band polarizers were in use for nuclear physics experiments at
width. MEOP is used for the Mainz target whereas at, MAMI. The rst one was the Toepler compressor which
e.g., Jeerson Lab the SEOP technique is applied. Both uses 17.6 kg mercury as a piston. The pressure achieved
techniques will be shortly explained and the advantages of in the 100 cm3 target cell was 1 bar and the polarization
each method discussed. It should be mentioned that there could be increased from 38% to 49% from 1993 to 1995.
is a third method to polarize 3 He. Here high magnetic The target cell was lled in a continuous ow (0.1 bar l/h)
elds and low temperatures are needed which leads to po- with polarized gas and the polarization loss from the low
larizations of 38% in solid 3 He [17]. Due to the low heat pressure pumping cell to the target was 30%. The increase
conductivity of the solid 3 He this method is not suitable of the polarization was achieved by coating the target cell
for nuclear-physics experiments with electron beams. with cesium to reduce the relaxation of the polarization
In MEOP as well as in SEOP angular momentum is due to collisions with the container material (glass). The
transferred to the atomic electrons by resonant absorption Toepler compressor was developed for the rst measure-
of circularly polarized light and subsequent re-emission of ment of the electric form factor of the neutron Gen which
unpolarized light. A magnetic eld of 530 G denes the is described below. Nowadays the compression stage is re-
quantization axis. In MEOP an atomic transition in 3 He placed by one titanium piston which allows a production
is pumped whose lower level is the metastable 23 S1 state. rate of 1.5 bar l/h. The polarization losses are negligible
It is reached by a weak gas discharge (a fraction of 106 and the target cell is lled with 5 bar and 75% polariza-
atoms is excited). Therefore this method works only at low tion. This is a great improvement and increases the per-
pressures of about 1 mbar which also guarantees a su- formance of the nuclear physics experiment signicantly.
ciently long lifetime of the 23 S1 state. With moderate laser
power of about 1020 W and for large gas quantities of 20
liter at 1 mbar a polarization up to 80% can be reached in
a minute. The formerly used LNA-Laser ( = 1083 nm,
3 Experiments with polarized 3 He
10 W) is nowadays replaced by two ytterbium ber
3.1 The electric form factor of the neutron
lasers (15 W each). Due to hyperne coupling the elec-
tronic polarization results in a corresponding alignment 3.1.1 Motivation and techniques
of the nuclear spin. Subsequent collisions between polar-
ized 3 He -atoms in the rst excited metastable state and Form factors describe the contribution from the inner
unpolarized 3 He-atoms in the ground state transfer the structure of a scatterer to the interaction. For spin-1/2
nuclear polarization to the ground state 3 He. The process particles there are two form factors determining the
of metastable-exchange collisions is fast and has a large electromagnetic response, the magnetic and the electric
cross section (1015 cm2 ). Therefore this method is quite form factor. They are related via a Fourier transforma-
ecient. The drawback of this method is that it can only tion to the magnetic and to the charge distribution (see
be applied at low pressures of about 1 mbar. sect. 3.1.5), respectively. A form factor independent of
In SEOP an alkali-vapor (usually Rb) is optically the momentum transfer q to the particle would indicate
pumped by the circularly polarized light provided by a Ti- a point-like distribution, hence any q-dependence points
sapphire laser or by diode lasers tuned to the D1 -resonance to an underlying substructure. The electric form factor
line of 795 nm. Once the Rb is polarized, the polarization Gen of the neutron is particularly sensitive to its internal
is transferred to the 3 He via spin-exchange collisions. The structure because it is not obscured by the total charge
spin-exchange mechanism proceeds via the hyperne in- as in the case for the proton. The substructure of the nu-
teraction between the 3 He nucleus and the Rb valence cleon is determined by the (sea-)quarks and the gluons.
electron. This can induce both species to ip their spin.
1
Because this interaction is weak the cross section for this In the Hermes experiment the cell was cooled down to 25 K
process is small (1024 cm2 ). Therefore optically thick Rb to achieve a compression factor of 3.5.
A1 and A3 Collaboration (D. Rohe): Experiments with polarized 3 He at MAMI 31
0.04
0.02 3
3.1.2 Form factor measurements in the experimental hall He(e,en)
A3 at MAMI 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
2 2
Q (GeV/c)
The setup for the Gen experiment in the A3 experimen-
Fig. 3. (Uncorrected) result from the rst measurement of Gen
tal hall is shown in g. 2. Electrons were detected in
[7]. Two double-polarization experiments
a calorimeter of 256 closely packed lead glass counters at MAMI using 3 He
performed at Bates in the same time period are also shown [24,
(4 4 cm) with a solid angle of 100 msr in a distance
25, 26]. For comparison the data already shown in g. 1 are
of 1.9 m. The shower produced by an electron extends
displayed as well.
over about 10 modules. Therefore the energy summed
over clusters of detectors was used leading to an energy
resolution E/E = 20% FWHM. This moderate energy
In this setup the already mentioned Toepler compres-
resolution was sucient to separate the inelastic contri- with polarizations in-
bution from the quasielastically scattered events. The in- sor was used to produce 1 bar 3 He
elastic events, mainly resulting from -production in the creasing from 38% to 49% from 1993 to 1995. At the same
-resonance, have vanishing asymmetry. In case of an ad- time the electron polarization could be increased from 30
mixture this would dilute the asymmetry. In front of the to 50% by changing the cathode from a bulk to a strained
calorimeter a focusing air Cerenkov detector was placed layer GaAsP. Keeping in mind that the statistical error of
which suppresses background from electrons scattered on the asymmetry decreases with (Pe PT T )1 (T : measure-
the exit or entrance windows of the target cell or the beam ment time) both improvements enhance the performance
line. Further it serves to discriminate photons and pions of the experiment signicantly.
from electrons. With this setup Gen was measured at Q2 =
The neutrons were detected in a plastic scintillator ar- 0.31 (GeV/c)2 [7,8]. The pilot experiment of Meyerho et
ray which covers 250 msr and therefore the entire Fermi al. in 1993 [7] did only use a quarter of the detector setup
cone. It consisted of two walls and could also be used as shown in g. 2. Its result is shown in g. 3 together with
a neutron polarimeter for the Gen -measurement using the other double-polarization experiments performed at Bates
reaction D(e, en) [22,23]. The overall detector thickness e, e ) [24,25] and D(e, en) [26].
at the same time using 3 He(
of 40 cm yields a neutron detection eciency of n = 32%. In g. 3 the uncorrected results are shown. In the mid of
The neutron detector was shielded with 5 cm lead on the the 90ies it was not clear that the measured value for Gen
front and surrounded by 1 m concrete against electromag- as polarized neutron target needs a large cor-
using 3 He
netic background. rection accounting for FSI. No exact Faddeev calculation
A1 and A3 Collaboration (D. Rohe): Experiments with polarized 3 He at MAMI 33
solid line represents a Monte Carlo simulation using sim- 0.08 D(e,en)
ple kinematical relations valid in Born approximation. The 0.07
D(e,en)
3
good agreement supports that the contribution from non- He(e,en)
0.06
quasielastic events is negligible.
0.05
The hadron detector was placed in the direction of q.
Gen
It consists of four layers with ve scintillator bars each. In 0.04
front of the detector two layers of E detectors discrim- 0.03
inate protons and neutrons. In 160 cm distance from the 0.02
target the detector covers a solid angle of about 100 msr.
0.01
The entire detector was shielded with 10 cm lead except
for an opening towards the target where a reduced shield 0
0 0.5 1 1.5 2
of 2 cm was used. In addition a lead collimator in front 2
Q (GeV/c)
2
Fig. 8. Experimental results of A (left) and A (right) for the central region of the quasi-elastic peak as a function of the
scattering angle of the knocked-out proton. The result of the full (PWIA) calculation is shown with solid (dashed) line. The
results of three full calculations, however with a non-relativistic current (dot), or with only a (v/c)2 correction (dot-dot-dashed),
or with non-relativistic kinematics (dot-dash) are also shown.
in the hadron detector and the scintillator plane of spec- ular, antiparallel and antiperpendicular asymmetry alter-
trometer A, was negligible. In order to study the eect nately with the purpose to reduce the systematic errors.
of FSI on the asymmetries in dierent kinematic regions, The target cell was of the same kind as already used for
the quasi-elastic peak is divided into two regions of . One the Gen measurement (see sect. 3.1.4). A new polarizer
region covers the peak and therefore emphasizes low nu- was used consisting of one-stage titanium compressor [45]
cleon momenta whereas the other region covers the low with eectively no polarization loss during the transfer
tail sensitive preferentially to high nucleon momenta. The from the low pressure gas reservoir to the target container.
events in each of the two regions are summed over the en- The 3 He was optically pumped with two Ytterbium ber
tire acceptance of the out-of-plane angle of electron and lasers each providing 15 W on the resonance transition
proton and over the electron scattering angle in a range (1083 nm). With this setup an initial target polarization
from 75.8 to 81.8 . of 70 to 75% could be achieved. Averaged over the beam
In g. 8 the parallel and perpendicular asymmetries time period and accounting for relaxation a target polar-
in the central region of the quasielastic peak are shown ization PT of (49.8 0.3(stat.) 2(syst.))% was obtained.
as a function of the scattering angle of the proton. They From the measured kinematic variables in the two
are compared to the theory which contains the two-body spectrometers, the missing energy is reconstructed accord-
(2BB) and three-body breakup (3BB). The 3BB channel ing to
is integrated over the rst 26 MeV. As can be seen from Em = E Ee Tp TR . (10)
the gures the PWIA calculation (dashed line) clearly dis-
agrees with the data. From the calculations which include Here, E (Ee ) is the initial (nal) electron energy and Tp
FSI23 only, the one with non-relativistic kinematics (dot- is the kinetic energy of the outgoing proton. TR is the ki-
dashed line) cannot describe the experimental results. Rel- netic energy of the (undetected) recoiling (A1)-system,
ativistic (solid line) or non-relativistic (dots) treatment which is reconstructed from the missing momentum un-
of the current operator does not make a large dierence. der the assumption of 2BB. The resulting Em distribution
The calculation taking into account relativistic kinematics reconstructed from the data is shown in g. 9 as thick
and FSI23 provides a good description of the data. Both solid line. The resolution is limited mainly by the prop-
ingredients are important to achieve agreement with the erties of the target cell and not by the resolution of the
experimental results. spectrometers. The FWHM of 1 MeV allows a clear sep-
aration of the Em -regions where only 2BB or 2BB and
3BB contribute. The Em -region from 4.0 to 6.5 MeV is in-
3.3 Structure of 3 He terpreted as pure 2BB. This cut was chosen to avoid any
contribution from the 3BB-channel (starting at 7.7 MeV)
In the experiment described in the previous section it was considering the experimental Em resolution. In agreement
not possible to separate the 2BB and 3BB channel due with ref. [46], the yield of the 3BB is negligible beyond
to the limited resolution of the hadron detector. For a 25 MeV. Therefore the cut for the 3BB-channel was made
spectrome-
better understanding of the structure of 3 He from 7.5 to 25.5 MeV in the Em spectrum. Because the
ter B was taken for proton detection. The kinematics was 3BB resides on the radiation tail of the 2BB, the latter
limited to the central region of the quasielastic momen- has to be accounted for in the analysis of the 3BB-region
tum distribution at Q2 of 0.31 (GeV/c)2 . Each hour the of the measured spectrum. To this end, the tail was cal-
target spin was turned to measure the parallel, perpendic- culated in a Monte Carlo simulation which accounts for
A1 and A3 Collaboration (D. Rohe): Experiments with polarized 3 He at MAMI 37
2
x 10
0.2
counts per bin
1600
2BB
1400
1200 0 3BB
asymmetry
1000
800
600 -0.2
400 3BB
200
-0.4
0 || || ||
5 10 15 20 25 target spin direction
Em (MeV)
Fig. 10. Comparison of the data to the theoretical calcula-
Fig. 9. Experimental Em distribution (thick line) and the sim- tion for the 2BB and 3BB for the four target spin directions
ulation of the 2BB (thin red line). The dierence is shown as (anti)parallel (, ; left panel) and (anti)perpendicular (,
thick black line 3BB. ; right panel). In addition the combined sum for the parallel
and perpendicular position is shown ( and , respectively). To
facilitate the comparison, all 2BB (3BB)data are shown with
internal and external bremsstrahlung, ionization loss and
positive (negative) sign. PWIA: dot-dashed lines. Full calcu-
experimental energy resolution adjusted to the experimen- lation including FSI and MEC: solid lines. Statistical errors
tal distribution. The simulated 2BB distribution is shown point up, systematic uncertainties point down. For the 2BB
as thin red line in g. 9. Subtracting this from the data the size of the error bars is smaller than the symbols.
leads to the distribution belonging to the 3BB channel
which is also shown in g. 9.
The ratio of the Monte Carlo simulation of the 2BB to
A2BB . In the 2BB channel, the polarized 3 He-target can
the experimental data in the region of the 3BB is denoted thus be interpreted as a polarized proton target.
by a23 . For the region 7.5 < Em < 25.5 MeV it amounts to For the 3BB channel the situation is dierent. In
a23 = 0.434 0.002(stat.) 0.015(sys.). Then the asym- PWIA the asymmetry is almost zero for the 3BB which
metry A3BB for the 3BB-channel is extracted from the reects the fact that the two protons, which are domi-
asymmetry A2+3BB in the 3BB region by accounting for nantly in the S-state and thus have opposite spin orien-
the contribution from the radiation tail tation, now contribute equally to the knock-out reaction.
A2+3BB A2BB a23 The inclusion of FSI, however, leads to an asymmetry,
A3BB = . (11) which is larger and opposite in sign compared to the 2BB.
1 a23
The main eect comes from the np t-matrix (rescattering
All asymmetries are corrected for target and electron po- term). Since dierent spin combinations of the singlet and
larization. In g. 10 the parallel and perpendicular asym- target cannot be in-
triplet np t-matrix contribute, the 3 He
metries A3BB and A2BB are compared to two calculations terpreted as a polarized proton target in the 3BB channel.
of the Bochum-Krakow group. One uses PWIA only (dot-
dashed), the other accounts for full FSI and MEC (solid
line). The eect of MEC is negligible in this kinematics.
The data integrated over the total detector acceptance are 4 Summary and outlook
in good agreement with the calculation including FSI.
The calculation shows that the FSI contribution is In this contribution a review of the experiments with po-
small in the 2BB while it is large in 3BB. This suggests performed at MAMI was given. The eort to
larized 3 He
that the main contribution of FSI results from the rescat- build a machine to polarize 3 He started already in 1987.
tering term which does not exist in the 2BB, and not from The rst experiment at MAMI with 3 He was performed
direct FSI. This was also conrmed by further examina- in the experimental hall A3 to measure the electric form
tion of the theoretical result by Golak [47]. factor of the neutron. Experiments with the same purpose
In the 2BB channel the spins of the neutron and proton at higher Q2 followed, using improved target and detector
in the recoiling deuteron are coupled to one, therefore they setups in the three-spectrometer hall A1. With the new
are parallel. Consequently, in a simplied picture, the spin detector setup a better discrimination of inelastic events
of the second (knocked-out) proton must be antiparallel from the ones quasielastically scattered is possible. The
Correct
to the deuteron spin and thus to the spin of 3 He. performance of the target was steadily improved due to
coupling of the spins 1 and 1/2 to 1/2 leads to 33% polar- the development of new polarizers. This resulted in a more
ization of the knocked-out proton relative to that of the dense target (5 bar) with higher polarization (PT = 50%).
polarized 3 He-target. This is precisely what is observed as In addition the electron source was improved using a
38 The European Physical Journal A
strained layer crystal. This led to nowadays available cur- vice as well as Hartmuth Arenhovel, Hartmut Backe and Dieter
rents of 20 A with an electron polarization of 75%. Drechsel for the good atmosphere in the institute.
Parallel to the experiments the non-relativistic Fad-
deev calculation was developed by the Bochum-Krakow
group. One of the rst applications was the calculation References
of the correction of Gen at Q2 = 0.35 (GeV/c)2 due to 1. J. Golak et al., Phys. Rev. C 65, 064004 (2002).
FSI which leads to a deviation of the asymmetry mea- 2. R.W. Schulze, P.U. Sauer, Phys. Rev. C 48, 38 (1993).
from that for a free neutron. The inuence
sured on 3 He 3. B. Blankleider, R.M. Woloshyn, Phys. Rev. C 29, 538
of FSI was also conrmed by measuring the target asym- (1984).
metry Ay where the beam is unpolarized and the target 4. H. Gao et al., Phys. Rev. C 50, R546 (1994).
spin perpendicular to the scattering plane. This quantity 5. W. Xu et al., Phys. Rev. Lett. 85, 2900 (2000).
6. W. Xu et al., Phys. Rev. C 67, 012201(R) (2003).
is particularly sensitive to FSI and MEC contributions. 7. M. Meyerho et al., Phys. Lett. B 327, 201 (1994).
In PWIA it vanishes. Good agreement between data and 8. J. Becker et al., Eur. Phys. J. A 6, 329 (1999).
theory was found. 9. J. Golak et al., Phys. Rev. C 63, 034006 (2001).
Another experiment concentrated on the question 10. D. Rohe et al., Phys. Rev. Lett. 83, 4257 (1999).
when a relativistic calculation is needed and which ingre- 11. J. Bermuth et al., Phys. Lett. B 564, 199 (2003).
12. P. Achenbach et al., Eur. Phys. J. A 25, 177 (2005).
dients need to be treated relativistically. For this the reac-
13. C. Carasco et al., Phys. Lett. B 599, 41 (2003).
e, e p) was investigated at Q2 = 0.67 (GeV/c)2 .
tion 3 He( 14. H. Kamada, W. Glockle, J. Golak, Ch. Elster, Phys. Rev.
It turned out that the kinematics has to be treated rel- C 66, 044010 (2002).
ativistically already at this Q2 . On the other hand, a 15. G.K. Walters, F.D. Colgrove, L.D. Schearer, Phys. Rev.
relativistic current operator is much less important. At Lett. 8, 439 (1962).
present a relativistic calculation is only possible in PWIA 16. M.A. Bouchiat, T.R. Carver, C.M. Varnum, Phys. Rev.
and with FSI23 included. Lett. 5, 373 (1960).
17. D.G. Haase et al., Nucl. Instrum. Methods A 402, 341
To become more sensitive to the inner structure of 3 He (1998).
3
the 2BB and 3BB channels in the reaction He(e, e p) were 18. M. Gockeler et al., Nucl. Phys. A 755, 537 (2005).
separated. Also here the theoretical calculation is in good 19. S. Platchkov et al., Nucl. Phys. A 510, 740 (1990).
agreement with the data. It is interesting that in the 2BB 20. R. Schiavilla, I. Sick, Phys. Rev. C 64, 041002(R) (2001).
channel, which is almost unaected by FSI at the kine- 21. R.G. Arnold, C.E. Carlson, F. Gross, Phys. Rev. C 23, 363
matics of the present experiment, the 3 He target can be (1981).
22. C. Herberg et al., Eur. Phys. J. A 5, 131 (1999).
considered as a polarized proton target with the proton
23. M. Ostrick et al., Phys. Rev. Lett. 83, 276 (1999).
By contrast, the 3BB chan-
spin opposite to that of 3 He. 24. C.E. Jones-Woodward et al., Phys. Rev. C 44, 571 (1991).
nel is largely aected by FSI. In this case 3 He cannot be 25. A.K. Thompson et al., Phys. Rev. Lett. 68, 2901 (1992).
interpreted as polarized proton target. 26. T. Eden et al., Phys. Rev. C 50, R1749 (1994).
27. J.M. Laget, Phys. Lett. B 273, 367 (1991).
All these reactions considered so far were not sensi- 28. J.M. Laget, Phys. Lett. B 276, 398 (1992).
tive to MEC because the kinematics were chosen to cor- 29. H.E. Conzett, Nucl. Phys. A 628, 81 (1998).
respond to the top of the quasielastic peak and the Q2 30. H.R. Poolman, PhD Thesis, Vrije Universiteit te Amster-
was suciently high. At Q2 < 0.2 (GeV/c)2 MEC con- dam, 1999.
tribute signicantly to the reaction and modify the asym- 31. E. Wilms et al., Nucl. Instrum. Methods A 401, 491
metries. Since MEC are not so well understood as com- (1997).
pared to FSI it is planned to study kinematics which are 32. R. Surkau et al., Nucl. Instrum. Methods A 384, 444
sensitive to MEC. The data taken to measure Gen at (1997).
33. I. Passchier et al., Phys. Rev. Lett. 82, 4988 (1999).
Q2 = 0.25 (GeV/c)2 are aected by MEC in some kine- 34. H. Zhu et al., Phys. Rev. Lett. 87, 081801 (2001).
matical regions covered by the detector acceptance. 35. R. Madey et al., Phys. Rev. Lett. 91, 122002 (2003).
With the upgrade of MAMI to MAMI-C the Gen mea- 36. G. Warren et al., Phys. Rev. Lett. 92, 042301 (2004).
surement will be pushed to Q2 = 1.5 (GeV/c)2 . For this a 37. D.I. Glazier et al., Eur. Phys. J. A 24, 101 (2005).
new hadron detector is under construction which should 38. S. Galster et al., Nucl. Phys. B 32, 221 (1971).
have a higher neutron detection eciency. 39. J. Friedrich, Th. Walcher, Eur. Phys. J. A 17, 607 (2003).
40. H.-W. Hammer, D. Drechsel, Ulf-G. Meiner, Phys. Lett.
There are also plans to use polarized 3 He with (po- B 586, 291 (2004).
larized) photons in the A2 experimental hall. Then 3 He 41. R. Machleidt, F. Sammarruca, Y. Song, Phys. Rev. C 53,
would be used as a polarized neutron target to measure 1483 (1996).
42. R.B. Wiringa, V.G.J. Stoks, R. Schiavilla, Phys. Rev. C
the Gerasimov-Drell-Hearn sum rule. For this a new target
51, 38 (1995).
setup is needed which is already under consideration. 43. A. Deltuva et al., Phys. Rev. C 70, 034004 (2004).
44. H. Arenhovel, W. Leidemann, E. Tomusiak, Phys. Rev. C
46, 455 (1992).
Finally I want to thank Karl-Heinz Kaiser for the excellent
45. E.W. Otten, Europhys. News 35, 16 (2004).
beam quality at MAMI and for his eort to adjust and setup 46. R.E.J. Florizone et al., Phys. Rev. Lett. 83, 2308 (1999).
the beam for our sensitive experiments. Then I want to thank 47. J. Golak, private communication, 2005.
Jorg Friedrich and Thomas Walcher for their support and ad-
Eur. Phys. J. A 28, s01, 39 48 (2006)
DOI: 10.1140/epja/i2006-09-005-6 EPJ A direct
electronic only
M. Schwamba
Institut fur Kernphysik, Johannes Gutenberg-Universitat, D-55099 Mainz, Germany
/
Published online: 9 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. An overview over present achievements and future challenges in the eld of few-nucleon systems
is presented. Special emphasis is laid on the construction of a unied approach to hadronic and electro-
magnetic reactions on few-nucleon systems, necessary for studying the borderline between quark-gluon and
eective descriptions.
PACS. 13.40.-f Electromagnetic processes and properties 21.45.+v Few-body systems 25.30.-c Lepton-
induced reactions 25.20.-x Photonuclear reactions
Fig. 10. Results for deuteron electrodisintegration taken from [15]. Left panel: the structure functions fL , fT , fLT and fT T
for the kinematics of the NIKHEF experiment [27], i.e. Enp = 280 MeV, q 2 = 2.47 fm 2 . Notation of the curves: dashed: static
approach; full: retarded approach. The additional dash-dotted curves represent the results of the retarded approach where the
Coulomb monopole contribution of the recoil charge operator, depicted on the right, is switched o.
}effect G En
We now turn to a second example where the use of the
deuteron as an eective neutron target would be highly t N
N
t N
N
desirable. It deals with the investigation of the Gerasimov- d d T NN
Drell-Hearn sum rule (GDH) for various hadronic tar- N N
gets [32,33]. This sum rule links the anomalous magnetic
(a) (b)
moment of a particle to the energy weighted integral over
the spin asymmetry of the absorption cross section. In de- N
tail it reads for a particle of mass M , charge eQ, anoma-
t N
lous magnetic moment and spin S
N
d T
d P
2 2 e
2 N
I GDH = ( ) A
( ) = 4 S,
0 M2 (c)
(14)
Fig. 13. Considered diagrams for single pion production.
where P/A ( ) denote for a given photon momentum (a) impulse approximation (IA), (b) incorporation of N N -nal
the total absorption cross sections for circularly polarized state interaction (N N -FSI), (c) incorporation of N -nal state
photons on a target with spin parallel (P ) and antiparal- interaction (N -FSI).
lel (A) to the photon spin. This sum rule gives therefore a
very interesting relation between a ground state property
() of a particle and its whole excitation spectrum. Apart
from the general assumption that the integral in (14) con- sum rule the deuteron reaction cannot be considered just
verges, its derivation is based solely on rst principles like as an incoherent sum of the proton and the neutron re-
Lorentz and gauge invariance, unitarity, crossing symme- action. In order to obtain the small deuteron GDH value,
try and causality of the Compton scattering amplitude of strong anticorrelation eects between the dierent pos-
a particle. Consequently, a check for various targets, both sible channels for the deuteron must occur which are not
from the experimental as well as from the theoretical point present in the elementary case. This cancellation is a chal-
of view, would be very important. lenge for any theoretical framework since it requires the
Inserting the known anomalous magnetic moments of above-mentioned unied consistent treatment of hadronic
proton and neutron into (14), one obtains quite large GDH and electromagnetic properties for the dierent possible
sum rule values, i.e. IpGDH = 204.8 b for the proton and channels in a wide energy region.
InGDH = 233.2 b for the neutron. On the other side, the In the past years, considerable eorts have been un-
deuteron has a small anomalous magnetic moment d = dertaken in order to obtain a more quantitative under-
0.143 n.m. resulting in a very small GDH sum rule value standing of the GDH sum rule on the deuteron [35,36,37].
of IpGDH = 0.65 b. In the presently most sophisticated approach [37], besides
Whereas GDH measurements on proton targets can be deuteron photodisintegration also coherent and incoherent
directly performed (consider [34] and references therein), single and double pion production as well as -production
no free neutron target exists and one may try to extract are considered. At the moment, the aforementioned re-
InGDH from deuteron measurements. In contrast to the tarded approach is only available for the breakup channel.
extraction of GEn , this task is however much more com- Concerning incoherent single pion production, the consid-
plicated. First of all, let us recall that for the extraction ered mechanisms in our present realization are depicted
of the electric neutron form factor one specic reaction in g. 13. For the elementary production operator, the
(e.g. deuteron electrodisintegration) in one specic kine- MAID model [38] is used, allowing one to extend the cal-
matics (the quasi-free one) is su cient. On the other hand, culation up to photon energies of 1.5 GeV. Moreover, nal
concerning the GDH sum rule one has to determine total state interactions are perturbatively taken into account
inclusive cross sections, i.e. contributions in all possible up to the rst order in the corresponding N - and N N -
kinematics from very dierent reactions like scattering amplitudes. For coherent pion production, the
model of [39] is used taking into account pion rescattering
N N, N, N, . . . (15) by solving a system of coupled equations for the N N -
, N - and N N -channels. It is partially similar to our
for the nucleon, and approach discussed in section 2.2. However, no retarda-
tion concerning the N N -interaction and the correspond-
d N N, N N, d, N N, N N, . . . (16)
ing MEC is presently taken into account. For double-pion
for the deuteron have to taken into account. production the evaluation is based on a traditional eec-
These complications become even more serious if one tive Lagrangian approach similar to the one in [40]. It is
considers the sum of the proton and neutron value com- presented in great detail in [41].
pared to the deuteron value. If one assumes that the meson Although this treatment of the GDH sum rule on the
production on the deuteron is dominated by the quasi-free deuteron is presently the most sophisticated one, we are
production on the nucleons bound in the deuteron, one aware of specic shortcomings. The most serious one is the
would expect that IdGDH should be roughly IpGDH +InGDH . use of dierent approaches for the dierent reactions. In
This assumption is however wrong by more than two or- order to obtain a more unied picture, work is in progress
ders of magnitude. Consequently, concerning the GDH to adopt the discussed retarded approach not only to the
46 The European Physical Journal A
0
Bonn r-space
500
d 0np FSI eects are in general quite small. The same is true also
-100 retarded potential 400 d 0d
300
N 0N for -production. But FSI is nonnegligible for incoherent
() [b]
() [b]
-200
-300 d pn
200 neutral pion production due to the non-orthogonality of
the nal state wave in IA to the deuteron bound state wave
GDH
GDH
100
-400
Id
Id
0
-500
-100 d pn (x5)
function, see [42] for more details. Please note moreover
N N (x5)
-600 the signicant dierences between the deuteron and the
1 10 100 1000 500 1000 1500
[MeV] [MeV] corresponding nucleon values for I GDH (). This feature
0 20
occurs also in double-pion production where the largest
-20 0 contribution is coming from the + -channel. Here the
() [b]
() [b]
GDH
-40
Id
Id
-60
d -pp (IA+FSI) d +nn (IA+FSI)
d -pp (IA)
-
n p
-60 d +nn (IA)
+
p n
The contributions of various channels to the nite
-80
500 1000 1500 500 1000 1500
GDH integral (17) for nucleon and deuteron are listed in
[MeV] [MeV]
20
table 1. While for the neutron the total sum is about 8
150 - +
d 0 0 np (IA+FSI) d np (IA+FSI) % lower than the sum rule value, it is too large by about
- +
d 0 0 np (IA) d np (IA)
- +
N 0 0 N N N
28 % for the proton. Concerning the deuteron, each of
() [b]
() [b]
100
10
the dierent channels (apart from -production) produces
GDH
GDH
Id
() [b]
30 30
20 20
eral shortcomings as indicated above.
GDH
GDH
Id
10 10
Table 1. Contributions of various channels to the nite GDH integral (in b), integrated up to 0.8 GeV for photodisintegration,
1.5 GeV for single pion and -production and 2.2 GeV for double pion production on nucleon and deuteron, see [37] for further
details.
np Sum rule value
neutron 138.95 82.02 5.77 215.20 233.16
proton 176.38 93.93 8.77 261.54 204.78
deuteron 381.52 263.44 159.34 13.95 27.31 0.65
and 3 He at MAMI will lead to very stringent tests of our 20. H. Tanabe, K. Ohta, Phys. Rev. C 40, 1905 (1989).
present knowledge of nucleon and nuclear structure. Addi- 21. P. Wilhelm, H. Arenhovel, Phys. Lett. B 318, 410 (1993).
tional measurements are also desirable for electromagnetic 22. J. Arends et al., Nucl. Phys. A 412, 509 (1984).
reactions on more complex few-nucleon systems (A 4) 23. G. Blanpied et al., Phys. Rev. C 52, R455 (1995); Phys.
where nowadays for the rst time purely microscopic cal- Rev. C 61, 024604 (2000).
culations with the help of the Lorentz integral transform 24. R. Crawford et al., Nucl. Phys. A 603, 303 (1996).
method are possible. 25. W. Fabian, H. Arenhovel, Nucl. Phys. A 314, 253 (1979).
Summarizing, the study of few-nucleon systems is a 26. H. Arenhovel, W. Leidemann, E.L. Tomusiak, Phys. Rev.
very active eld both from the experimental as well as C 46, 455 (1992).
27. A. Pellegrino et al., Phys. Rev. Lett. 78, 4011 (1997).
theoretical point of view. The expected progress will be
28. M. Gari, H. Hyuga, Z. Phys. 277, 291 (1976).
very important for the future development of hadronic
29. H. Arenhovel, Czech. J. Phys. 43, 207 (1993).
physics in general. 30. H. Arenhovel, W. Leidemann, E.L. Tomusiak, Z. Phys. A
331, 123 (1988); 334, 363 (1989).
This is dedicated to the occasion of the retirement of 31. D.I. Glazier et al., Eur. Phys. J. A 24, 101 (2005).
H. Arenhovel, H. Backe, D. Drechsel, J. Friedrich, K-H. Kaiser 32. S.B. Gerasimov, Yad. Fiz. 2, 598 (1965) (Sov. J. Nucl.
and Th. Walcher. It has been supported by the Deutsche Phys. 2, 430 (1966)).
Forschungsgemeinschaft (SFB443). I would like to thank 33. S.D. Drell, A.C. Hearn, Phys. Rev. Lett. 16, 908 (1966).
H. Arenhovel for his careful reading of the manuscript and 34. H. Dutz et al., Phys. Rev. Lett. 93, 032003 (2004).
for various stimulating discussions. 35. H. Arenhovel, G. Kre , R. Schmidt, P. Wilhelm, Phys.
Lett. B 407, 1 (1997).
36. E.M. Darwish, H. Arenhovel, M. Schwamb, Eur. Phys. J.
A 17, 513 (2003).
References 37. H. Arenhovel, A. Fix, M. Schwamb, Phys. Rev. Lett. 93,
202301 (2004).
1. R. Machleidt, Adv. Nucl. Phys. 19, 189 (1989). 38. D. Drechsel, O. Hahnstein, S.S. Kamalow, L. Tiator, Nucl.
2. E.W. Schmid, H. Ziegelmann, The Quantum-Mechanical Phys. A 645, 145 (1999).
Three-Body Problem (Pergamon Press, Oxford and 39. P. Wilhelm, H. Arenhovel, Nucl. Phys. A 593, 435 (1995);
Friedrich Vieweg & Sohn, Braunschweig, 1974). 609, 469 (1996).
3. W. Glockle, The Quantum-Mechanical Few-Body Problem 40. J.A. Gomez Tejedor, E. Oset, Nucl. Phys. A 600, 413
(Springer Verlag, Berlin 1983). (1996).
4. R.B. Wiringa, V.G.J. Stoks, R. Schiavilla, Phys. Rev. C 41. A. Fix, H. Arenhovel, Eur. Phys. J. A 25, 115 (2005).
51, 38 (1995). 42. A. Fix, H. Arenhovel, Phys. Rev. C 72, 064005 (2005).
5. R. Machleidt, F. Sammarrucca, Y. Song, Phys. Rev. C 53, 43. J. Golak, R. Skibinski, H. Witala, W. Glockle, A. Nogga,
1483 (1996). H. Kamada, Phys. Rep. 415, 89 (2005).
6. R. Machleidt, Phys. Rev. C 63, 024001 (2001). 44. V.D. Efros, W. Leidemann, G. Orlandini, Phys. Lett. B
7. M. Schwamb, H. Arenhovel, P. Wilhelm, Th. Wilbois, 338, 130 (1994).
Phys. Lett. B 420, 255 (1998). 45. D. Andreasi, W. Leidemann, Ch. Reiss, M. Schwamb, Eur.
8. M. Schwamb, H. Arenhovel, Nucl. Phys. A 690, 647 Phys. J. A 24, 361 (2005).
(2001). 46. V.D. Efros, W. Leidemann, G. Orlandini, Phys. Rev. Lett.
9. M. Schwamb, H. Arenhovel, Nucl. Phys. A 690, 682 78, 432 (1997).
(2001). 47. V.D. Efros, W. Leidemann, G. Orlandini, E.L. Tomusiak,
10. M. Schwamb, H. Arenhovel, Nucl. Phys. A 696, 556 Phys. Rev. C 69, 044001 (2004).
(2001). 48. V.D. Efros, W. Leidemann, G. Orlandini, Phys. Rev. Lett.
11. Ch. Elster, W. Ferchlander, K. Holinde, D. Schutte, R. 78, 4015 (1997).
Machleidt, Phys. Rev. C 37, 1647 (1988). 49. S. Bacca, M. Marchisio, N. Barnea, W. Leidemann, G.
12. H. Popping, P.U. Sauer, X.-Z. Zhang, Nucl. Phys. A 474, Orlandini, Phys. Rev. Lett. 89, 052502 (2002).
557 (1987). 50. S. Bacca, H. Arenhovel, N. Barnea, W. Leidemann, G.
13. J. Haidenbauer, W. Plessas, Phys. Rev. C 30, 1822 (1984); Orlandini, Phys. Lett. B 603, 159 (2004).
Phys. Rev. C 32, 1424 (1985). 51. A. La Piana, W. Leidemann, Nucl. Phys. A 677, 423
14. S. Nozawa, B. Blankleider, T.-S.H. Lee, Nucl. Phys. A 513, (2000).
459 (1990). 52. S. Quaglioni, W. Leidemann, G. Orlandini, N. Barnea,
15. M. Schwamb, H. Arenhovel, Phys. Lett. B 588, 49 (2004). V.D. Efros, Phys. Rev. C 69, 044002 (2004).
16. R. Machleidt, D.R. Entem, J. Phys. G 31, S1235 (2005). 53. Ch. Reiss, W. Leidemann, G. Orlandini, E.L. Tomusiak,
17. E. Epelbaum, W. Glockle, U.-G. Meissner, Nucl. Phys. A Eur. Phys. J. A 17, 589 (2003).
747, 362 (2005). 54. Ch. Reiss, H. Arenhovel, M. Schwamb, Eur. Phys. J. A 25,
18. M. Walzl, U.-G. Meissner, Phys. Lett. B 513, 37 (2001). 171 (2005).
19. H. Krebs, V. Bernard, U.-G. Meissner, Eur. Phys. J. A 22, 55. D. Gazit, N. Barnea, Phys. Rev. C 70, 048801 (2004).
503 (2004).
Eur. Phys. J. A 28, s01, 49 57 (2006)
DOI: 10.1140/epja/i2006-09-006-5 EPJ A direct
electronic only
/
Published online: 11 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Dispersion relations provide a powerful tool to analyse the electromagnetic form factors of the
nucleon in both the space-like and the time-like regions with constraints from other experiments, unitarity,
and perturbative QCD. We give a brief introduction into dispersion theory for nucleon form factors and
present rst results from our ongoing form factor analysis. We also calculate the two-pion continuum
contribution to the isovector spectral functions drawing upon the new high statistics measurements of the
pion form factor by the CMD-2, KLOE, and SND collaborations.
PACS. 11.55.Fv Dispersion relations 13.40.Gp Electromagnetic form factors 14.20.Dh Protons and
neutrons
jem
where F (t) is a generic form factor. In the case of the
electric and Dirac form factors of the neutron, GnE and
F1n , the expansion starts with the term linear in t and the
p p normalization factor F (0) is dropped.
In this paper we give a brief introduction into disper- Based on unitarity and analyticity, dispersion relations
sion theory for nucleon form factors and present prelim- relate the real and imaginary parts of the electromagnetic
inary results from our ongoing form factor analysis. We (em) nucleon form factors. Let F (t) be a generic symbol
also calculate the two-pion continuum contribution to the for any one of the four independent nucleon form factors.
isovector spectral functions drawing upon the new high We write down an unsubtracted dispersion relation of the
statistics measurements of the pion form factor by the form
1 Im F (t )
CMD-2, KLOE, and SND collaborations. Finally we ad- F (t) = dt , (6)
dress the question of the range of the pion cloud and give t0 t t i
an outlook on future work. where t0 is the threshold of the lowest cut of F (t) (see
below) and the i denes the integral for values of t on the
2 Denitions cut.1 Equation (6) relates the em structure of the nucleon
to its absorptive behavior.
The electromagnetic (em) structure of the nucleon is de- The imaginary part Im F entering eq. (6) can be ob-
termined by the matrix element of the current operator tained from a spectral decomposition [13,14]. For this pur-
jem between nucleon states as illustrated in g. 1. pose it is most convenient to consider the em current ma-
Using Lorentz and gauge invariance, this matrix ele- trix element in the time-like region (t > 0), which is re-
ment can be expressed in terms of two form factors, lated to the space-like region (t < 0) via crossing symme-
try. The matrix element can be expressed as
em F2 (t)
p |j |p = u(p ) F1 (t) + i
q u(p), (1)
2M J = N (p)N (p)|jem (0)|0 (7)
where M is the nucleon mass and t = (p p) the four- 2
= u(p) F1 (t) + i
F2 (t)
(p + p) v(p),
momentum transfer. For data in the space-like region, it 2M
is often convenient to use the variable Q2 = t > 0. The
functions F1 (t) and F2 (t) are the Dirac and Pauli form where p and p are the momenta of the nucleon and anti-
factors, respectively. They are normalized at t = 0 as nucleon created by the current jem , respectively. The four-
momentum transfer in the time-like region is t = (p + p)2 .
F1p (0) = 1, F1n (0) = 0, F2p (0) = p , F2n (0) = n , (2)
Using the LSZ reduction formalism, the imaginary part
with p = 1.79 and n = 1.91 the anomalous magnetic of the form factors is obtained by inserting a complete set
moments of protons and neutrons in nuclear magnetons, of intermediate states as [13,14]
respectively.
It is convenient to work in the isospin basis and to Im J = (2)3/2 N p|JN (0)| (8)
decompose the form factors into isoscalar and isovector Z
parts, |jem (0)|0 v(p) 4 (p + p p ),
1 p 1 p
Fis = (F + Fin ), Fiv = (F Fin ), (3) where N is a nucleon spinor normalization factor, Z is
2 i 2 i
the nucleon wave function renormalization, and JN (x) =
where i = 1, 2.
J (x)0 with JN (x) a nucleon source. This decomposition
The experimental data are usually given for the Sachs
is illustrated in g. 2. It relates the spectral function to
form factors
on-shell matrix elements of other processes.
GE (t) = F1 (t) F2 (t), (4) The states | are asymptotic states of momentum p
GM (t) = F1 (t) + F2 (t), which are stable with respect to the strong interaction.
They must carry the same quantum numbers as the cur-
where = t/(4M 2 ). In the Breit frame, GE and GM may rent jem : I G (J P C ) = 0 (1 ) for the isoscalar current
be interpreted as the Fourier transforms of the charge and and I G (J P C ) = 1+ (1 ) for the isovector component of
magnetization distributions, respectively.
1
The nucleon radii can be dened from the low-t expan- The convergence of an unsubtracted dispersion relation for
sion of the form factors, the form factors has been assumed. We could have used a once
subtracted dispersion relation as well since the normalization
F (t) = F (0) 1 + tr2 /6 + . . . , (5) of the form factors is known.
H.-W. Hammer: Nucleon form factors in dispersion theory 51
N 60
KLOE
CMD-2
50 SND
40
40
2
|F (t)|
jem
30 30
N 20
Fig. 2. The spectral decomposition of the nucleon matrix el- 0.5 0.6
ement of the electromagnetic current jem . 10
0.06 Im F iS Im F iV
spectral function [1/M ]
4
2
2ImGE/t
S
2
2ImGM/t
0.04
S
t t
, KK
0.02
Fig. 5. Illustration of the spectral function used in the disper-
sion analysis. The vertical dashed line separates the well-known
low-mass contributions (2, KK, and continua as well as
0 the pole) from the eective poles at higher momentum trans-
0 20 40
2 fers.
t [M ]
Fig. 4. The 2 spectral function using the new high statistics 5 Spectral functions
data for the pion form factor [39, 40, 41]. The spectral functions
weighted by 1/t2 are shown for GE (solid line) and GM (dash- As discussed above the spectral function can at present
dotted line) in units of 1/M4 . The previous results by Hohler only be obtained from unitarity arguments for the lightest
et al. [42] (without - mixing) are shown for comparison by two-particle intermediate states (2 and KK) [19,36,37].
the green lines. The continuum contribution can be obtained from the
Bonn-Julich model [46].
The remaining contributions can be parametrized by
line) and GM (dash-dotted line). The previous results by
vector meson poles. On one hand, the lower mass poles can
Hohler et al. [42] (without - mixing) are given for com-
be identied with physical vector mesons such as the and
parison by the gray/green lines. The general structure of
the . In the the case of the 3 continuum, e.g., it has been
the two evaluations is the same, but there is a dierence
shown in ChPT that the nonresonant contribution is very
in magnitude of about 10%. The dierence between the
small and the spectral function is dominated by the [25].
three data sets for the pion form factor is very small and
The higher mass poles on the other hand, are simply an
indicated by the line thickness. The dierence in the form
eective way to parametrize higher mass strength in the
factors is largest in the -peak region (cf. g. 3), but this
spectral function.
region is suppressed by the N N amplitudes f 1
(t)
For our current best t, the spectral function includes
which show a strong fall-o as t increases.
the 2, KK, and continua from unitarity and the
The spectral functions have two distinct features.
pole. In addition to that there are a number of eective
First, as already pointed out in [15], they contain the
poles at higher momentum transfers in both the isoscalar
important contribution of the -meson with its peak at
and isovector channels. The spectral function then has the
t 30M2 . Second, on the left shoulder of the , the
general structure
isovector spectral functions display a very pronounced en-
hancement close to the two-pion threshold. This is due to
the logarithmic singularity on the second Riemann sheet Im Fis (t) = Im FiK K (t) + Im Fi (t)
located at tc = 4M2 M4 /M 2 = 3.98M2 , very close to + aVi (MV2 t), i = 1, 2, (11)
the threshold. This pole comes from the projection of the V =,s1 ,...
nucleon Born graphs, or in modern language, from the
Im Fiv (t) = Im Fi2 (t)
triangle diagram.
If one were to neglect this important unitarity correc- + aVi (MV2 t), i = 1, 2, (12)
tion, one would severely underestimate the nucleon isovec- V =v1 ,...
tor radii [17],
which is illustrated in in g. 5. The vertical dashed line
6 dt separates the well-known low-mass contributions to the
r2 vi = Im Gvi (t), (10)
4M2 t2 spectral function from the eective poles at higher mo-
mentum transfers.
where i = E, M . In fact, precisely the same eect is ob- In our ts, we also include the widths of the vector
tained at leading one-loop accuracy in relativistic chiral mesons. The width and mass of the are taken from the
perturbation theory [44,45]. This topic was also discussed particle data tables while the masses and widths of the
in heavy baryon ChPT [25,27] and in a covariant calcula- eective poles are tted to the form factor data. We have
tion based on infrared regularization [26]. Thus, the most performed various ts with dierent numbers of eective
important 2 contribution to the nucleon form factors can poles and including/excluding some of the continuum con-
be determined by using either unitarity or ChPT (in the tributions. In sect. 7, we will discuss preliminary results
latter case, of course, the contribution is not included). of this ongoing eort.
H.-W. Hammer: Nucleon form factors in dispersion theory 53
with n = 0 for F1 and n = 0, 1 for F2 . The asymptotic This work Ref. [23] Recent determ.
p
behavior of eq. (13) is obtained by choosing the residues of rE [fm] 0.84...0.857 0.848 0.886(15) [53, 54, 55]
p
the vector meson pole terms such that the leading terms rM [fm] 0.85...0.875 0.857 0.855(35) [54, 56]
in the 1/t-expansion cancel. n
rE [fm] 0.12...0.10 0.12 0.115(4) [48]
n
The logarithmic term in eq. (13) was included in some rM [fm] 0.86...0.88 0.879 0.873(11) [57]
of our earlier analyses [20,21,23] but has little impact
on the t. The particular way this constraint was imple-
In Table 1, we give the nucleon radii extracted from
mented, however, lead to an unphysical logarithmic singu-
our t. The neutron radius is included as a soft constraint
larity of the form factors in the time-like region. In order
in our t and therefore not a prediction.3 The other nu-
to be able to include the data for the form factors at large
cleon radii are generally in good agreement with other
time-like momentum transfers, the logarithmic constraint
recent determinations using only low-momentum-transfer
is not enforced in the current analysis.
data given in the table. Our result for the proton radius,
The number of eective poles in eqs. (11, 12) is deter- however, is somewhat small. This was already the case in
mined by the stability criterion discussed in detail in [50]. the dispersion analyses of refs. [20,23]. We speculate that
In short, we take the minimum number of poles necessary the reason for this discrepancy lies in inconsistencies in the
to t the data. For the preliminary results discussed in the data sets. In this type of global analysis all four form fac-
next section, we took 4 eective isoscalar poles and 3 ef- tors are analyzed simultaneously and both data at small
fective isovector poles whose residua, masses, and widths and large momentum transfers enter. This can be an ad-
are tted to the data. The number of free parameters is vantage or a disadvantage depending on the question at
strongly reduced by the various constraints (unitarity, nor- hand. Another possible reason for the discrepancy is 2
malizations, superconvergence relations), so that we end physics which was neglected in the data analysis of most
up with 19 free parameters in the preliminary t presented older experiments [58].
in the next section. Our general strategy is to reduce the
3
number of parameters even further without sacricing the A soft constraint is not implemented exactly but deviations
quality of the t. from the constraint are penalized in the 2 of the t.
54 The European Physical Journal A
0.1 1.2
0.08
GM /(nGD)
1
0.06
n
GE
n
0.04
0.8
0.02
0 0.6
0 0.5 1 1.5 2 0.01 0.1 1 10
2 2 2 2
Q [GeV ] Q [GeV ]
1.5 1.2
GM /(pGD)
1 1
GE /GD
p
0.5 0.8
0 0.6
0.01 0.1 1 10 0.01 0.1 1 10 100
2 2 2 2
Q [GeV ] Q [GeV ]
p p
Fig. 6. The nucleon electromagnetic form factors for space-like momentum transfer. The results for G n M , GE , GM are normalized
to the dipole t. The dash-dotted line gives the result of ref. [23], while the the solid line indicates our preliminary best t.
8 Time-like data from the pioneering FENICE experiment [11]. They have
been analyzed under both the assumption |GE | = |GM |
We have also performed rst ts that include data in and |GE | = 0. The latter hypothesis is favored by the mea-
the time-like region. The extraction of these data is more sured angular distributions [11]. Neither data set could be
challenging than in the space-like region. At the nucleon- described by the analysis [22].
antinucleon threshold, the electric and magnetic form fac- In g. 7, we show the current status of the analysis of
tors are equal by denition: GM (4M 2 ) = GE (4M 2 ), while the time-like data for the magnetic form factors. For the
one expects the magnetic form factor to dominate at large proton magnetic form factor, data up to momentum trans-
momentum transfer. Moreover, the form factors are com- fers t 15 GeV2 have been included [6,7,8,9,10]. Our pre-
plex in the time-like region, since several physical thresh- liminary t gives a good description in the threshold re-
olds are open. Separating |GM | and |GE | unambiguously gion but starts to deviate signicantly around t 5 GeV 2 .
from the data requires a measurement of the angular dis- The data for t 10 GeV2 are well described. This seems
tribution, which is di cult. In most experiments, it has to be due to a slight inconsistency in the data around
been assumed that either |GM | = |GE | (which should be 5 GeV2 and for t 10 GeV2 . This question deserves fur-
a good approximation close to the two-nucleon threshold) ther attention.
or |GE | = 0 (which should be a good approximation for The status for the neutron form factor is the same as
large momentum transfers). Most recent data have been in the previous analysis [22]: Neither of the two data sets
presented using the latter hypothesis. from ref. [11] can be described. Even though we are not
The time-like data were already included in the disper- yet in the region where perturbative QCD is applicable, it
sion analyses of refs. [21,22]. The proton magnetic form comes as a surprise that the neutron form factor is larger
factor up to t 6 GeV2 was well described by these anal- in magnitude than the proton one. Perturbative QCD pre-
yses. Data at higher momentum transfers were not in- dicts asymptotically equal magnitudes. In any case, there
cluded. The data for the neutron magnetic form factor are is interesting physics in the time-like nucleon form factors
H.-W. Hammer: Nucleon form factors in dispersion theory 55
p 1.2
|GM (t)| v
GM
4r (r) [1/fm]
0.4
v
0.8 GE
2
0.2
0.4
0
5 10 15
2 0.0
t [GeV ] 0.0 0.5 1.0 1.5 2.0
0.8 r [fm]
n Fig. 8. Pion cloud contribution to the nucleon charge density.
|GM (t)| The lines show the result of Friedrich and Walcher [34], while
0.6
the bands give the result of ref. [35]. Only the long-range con-
tributions for r > 1 fm are meaningful for the comparison of
the two results.
0.4
Friedrich and Walcher (FW) recently analysed the em nu- where i = E, M . The 2 contribution from t 40M2 is
cleon form factors and performed various phenomenolog- small and can be neglected [35].
ical ts [34]. Their ts showed a pronounced bump struc- The result for the pion cloud contribution to the nu-
ture in GnE which they interpreted using an ansatz for cleon charge density is shown in g. 8. The lines show
the pion cloud based on the idea that the proton can be the result of FW [34], while the bands give the result of
ref. [35]. Only the long-range contributions for r > 1 fm
thought of as a virtual neutron-positively charged pion
pair. They found a very long-range contribution to the should be compared since the separation of the short-
charge distribution in the Breit frame extending out to range part into resonant and nonresonant contributions
about 2 fm which they attributed to the pion cloud. While is arbitrary. In comparison with ref. [34], the 2 contin-
naively the pion Compton wave length is of this size, these uum contribution to the charge density is generally much
ndings are indeed surprising if compared with the pion smaller at distances beyond 1 fm, e.g., by a factor of 3 for
cloud contribution due to the 2 continuum contribution vE (r) at r = 1.5 fm. We emphasize that this result is ob-
to the isovector spectral functions discussed in sect. 4. tained from independent physical information that deter-
mines the 2 continuum (pion form factor and N N
As was shown by Hammer, Drechsel, and Mei ner [35],
amplitudes, cf. sect. 4) and not from form factor ts.
these latter contributions to the long-range part of the
As a consequence, it remains to be shown how the pro-
nucleon structure are much more conned in coordinate
posed long-range pion cloud can be reconciled with what
space and agree well with earlier (but less systematic) cal-
culations based on chiral soliton models, see, e.g., [59]. 4
Note that this separation is not unique. It is only meaning-
In the dispersion-theoretical framework, the longest-range ful for the long-range part. The separation of the short-range
part of the pion cloud contribution to the nucleon form part is model- and even representation-dependent.
56 The European Physical Journal A
0.05
10 Summary & outlook
/
Published online: 12 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Chiral perturbation theory is the eective eld theory of the strong interactions at low energies.
We will give a short introduction to chiral perturbation theory for mesons and will discuss, as an example,
the electromagnetic polarizabilities of the pion. These have recently been extracted from an experiment
on radiative + photoproduction from the proton (p + n) at the Mainz Microtron MAMI. Next we
will turn to the one-baryon sector of chiral perturbation theory and will address the issue of a consistent
power counting scheme. As examples of the heavy-baryon framework we will comment on the extraction of
the axial radius from pion electroproduction and will discuss the generalized polarizabilities of the proton.
Finally, we will discuss two recently proposed manifestly Lorentz-invariant renormalization schemes and
illustrate their application in a calculation of the nucleon electromagnetic form factors.
PACS. 11.10.Gh Renormalization 11.30.Rd Chiral symmetries 13.40.-f Electromagnetic processes and
properties 13.40.Gp Electromagnetic form factors 13.60.Fz Elastic and Compton scattering 13.60.Le
Meson production
a simultaneous expansion of matrix elements in 1/m and is a result of the nite quark masses of the u, d and s
1/(4F ). However, there is price one pays when giving up quarks. This explicit symmetry breaking in terms of the
manifest Lorentz invariance of the Lagrangian. At higher quark masses is treated perturbatively.
orders in the chiral expansion, the expressions due to the The symmetries as well as the symmetry breaking pat-
1/m corrections of the Lagrangian become increasingly tern of QCD once the quark masses are included are
complicated [11,12]. Moreover, not all of the scattering mapped onto the most general (eective) Lagrangian for
amplitudes, evaluated perturbatively in the heavy-baryon the interaction of the Goldstone bosons. The Lagrangian
framework, show the correct analytical behavior in the is organized in the number of the (covariant) derivatives
low-energy region [13]. In recent years, there has been and of the quark mass terms [1,2,3,25,26,27,28,29,30,31]
a considerable eort in devising renormalization schemes
leading to a simple and consistent power counting for the L = L2 + L4 + L6 + , (2)
renormalized diagrams of a manifestly Lorentz-invariant 1
where the lowest-order Lagrangian is given by
approach [14,15,16,17,18,19,20,21].
In the following we will highlight a few topics in chiral F2
perturbation theory which have been subject of experi- L2 = Tr D U (D U ) + U + U . (3)
4
mental tests at the Mainz Microtron MAMI.
Here,
0
+
2 Chiral perturbation theory for mesons U (x) = exp i , = 20 ,
F 2
2.1 The eective Lagrangian and Weinbergs power
counting scheme is a unimodular unitary (2 2) matrix containing the
Goldstone boson elds. In eq. (3), F denotes the pion-
decay constant in the chiral limit: F = F [1 + O(m)] =
The starting point of mesonic chiral perturbation theory
92.4 MeV. When including the electromagnetic interac-
is a chiral SU (Nl )L SU (Nl )R U (1)V symmetry of the
tion, the covariant derivative is dened as D U = U +
QCD Lagrangian for Nl massless (light) quarks:
ieA [Q, U ], where Q = diag(2/3, 1/3) denotes the quark
charge matrix. We work in the isospin-symmetric limit
Nl
1
L0QCD = (qR,l iD / qL,l ) G,a Ga . (1)
/ qR,l + qL,l iD mu = md = m. The quark masses are contained in
l=1
4 = 2Bm = M 2 , where M 2 denotes the lowest-order ex-
pression for the squared pion mass and B is related to the
In eq. (1), qL,l and qR,l denote the left- and right-handed quark condensate qq 0 in the chiral limit. The next-to-
components of the light quark elds. Here, we will be con- leading-order Lagrangian contains 7 low-energy constants
cerned with the cases Nl = 2 and Nl = 3 referring to li [2]
massless u and d or u, d and s quarks, respectively. Fur-
thermore, we will neglect the terms involving the heavy- 1 L R
L4 = l5 Tr(f U fL U ) Tr(f fL + f fR )
R
quark elds. The covariant derivative D qL/R,l contains 2
the avor-independent coupling to the eight gluon gauge l6 R
potentials, and G,a are the corresponding eld strengths. +i Tr f D U (D U ) + f
L
(D U ) D U + , (4)
2
The Lagrangian of eq. (1) is invariant under separate
global SU (Nl )L/R transformations of the left- and right- where we have displayed those terms which will be relevant
handed elds. In addition, it has an overall U (1)V sym- for the discussion of Compton scattering below. In that
metry. Several empirical facts give rise to the assumption case, the eld strength is given by
that this chiral symmetry is spontaneously broken down
to its vectorial subgroup SU (Nl )V U (1)V . For example,
R
f L
= f = e( A A )Q.
the low-energy hadron spectrum seems to follow multiplic- In addition to the most general Lagrangian, one needs
ities of the irreducible representations of the group SU (Nl ) a method to assess the importance of various diagrams cal-
(isospin SU (2) or avor SU (3), respectively) rather than culated from the eective Lagrangian. Using Weinbergs
SU (Nl )L SU (Nl )R , as indicated by the absence of de- power counting scheme [1] one may analyze the behavior
generate multiplets of opposite parity. Moreover, the light- of a given diagram calculated in the framework of eq. (2)
est mesons form a pseudoscalar octet with masses that under a linear re-scaling of all external momenta, pi tpi ,
are considerably smaller than those of the corresponding and a quadratic re-scaling of the light quark masses,
vector mesons. According to Colemans theorem [22], the m t2 m, which, in terms of the Goldstone boson masses,
symmetry pattern of the spectrum reects the invariance corresponds to M 2 t2 M 2 . The chiral dimension D of a
of the vacuum state. Therefore, as a result of Goldstones given diagram with amplitude M(pi , m) is dened by
theorem [23,24], one would expect 6 3 = 3 or 16 8 = 8
massless Goldstone bosons for Nl = 2 and Nl = 3, respec- M(tpi , t2 m) = tD M(pi , m), (5)
tively. These Goldstone bosons have vanishing interactions
1
as their energies tend to zero. Of course, in the real world, In the following, we will give equations for the two-avor
the pseudoscalar meson multiplet is not massless which case.
S. Scherer: Chiral perturbation theory 61
where, in n dimensions,
D = nNL 2I +
2kN2k (6)
k=1
2
Fig. 1. One-loop contribution to the pion self-energy. The
= 2 + (n 2)NL + 2(k 1)N2k
(7) number 2 in the interaction blob refers to L2 .
k=1
2 in 4 dimensions.
e2 1 l6 l 5
+ = + = 2
4 (4F )2 M 6 p n
= (2.64 0.09) 104 fm3 , on a pion as a sub diagram in the t channel, where t = (pn
pp )2 .
which relates Compton scattering on a charged pion,
+ + , in terms of a chiral Ward identity to radiative
charged-pion beta decay, + e+ e . The linear combi-
pion-nucleus bremsstrahlung, Z Z [41], radia-
nation l6 l5 of scale-independent low-energy constants [2]
tive pion photoproduction on the nucleon, p + n [42,
is xed using the most recent determination of the ratio of 43], and pion pair production in e+ e scattering, e+ e
the pion axial-vector form factor FA and the vector form
e+ e + [44,45,46,47]. The results of the older experi-
factor FV via the radiative pion beta decay [35]:
ments are summarized in table 1.
1 FA The potential of studying the inuence of the pion po-
= (l6 l5 ) = = 0.443 0.015. larizabilities on radiative pion photoproduction from the
6 FV
proton was extensively studied in [48]. In terms of Feyn-
A two-loop analysis (O(q 6 )) of the charged-pion polariz- man diagrams, the reaction p + n contains real
abilities has been worked out in [36,37]2 : Compton scattering on a charged pion as a pion pole dia-
gram (see g. 3). In the recent experiment on p + n
( + )+ = (0.3 0.1) 104 fm3 , (11) at the Mainz Microtron MAMI [43], the cross section was
( )+ = (4.4 1.0) 104 fm3 . (12) obtained in the kinematic region 537 MeV < E < 817
MeV, 140 cm
180 . The values of the pion polar-
The degeneracy + = + is lifted at the two-loop level. izabilities have been obtained from a t of the cross sec-
The corresponding corrections amount to an 11% (22%) tion calculated by dierent theoretical models to the data
change of the O(q 4 ) result for + (+ ), indicating a simi- rather than performing an extrapolation to the t-channel
lar rate of convergence as for the -scattering lengths [2, pole of the Chew-Low type [49,50]. Figure 4 shows the
38]. The eect of the new low-energy constants appear- experimental data, averaged over the full photon beam
ing at O(q 6 ) on the pion polarizability was estimated via energy interval and over the squared pion-photon center-
resonance saturation by including vector and axial-vector of-mass energy s1 from 1.5 M2 to 5 M2 as a function of
mesons. The contribution was found to be about 50% of the squared pion momentum transfer t in units of M2 .
the two-loop result. However, one has to keep in mind For such small values of s1 , the dierential cross section
that [36,37] could not yet make use of the improved anal- is expected to be insensitive to the pion polarizabilities.
ysis of radiative pion decay which, in the meantime, has Also shown are two model calculations: model 1 (solid
also been evaluated at two-loop accuracy [39,40]. Taking curve) is a simple Born approximation using the pseu-
higher orders in the quark mass expansion into account, doscalar pion-nucleon interaction including the anomalous
Bijnens and Talavera obtain (l6 l5 ) = 2.98 0.33 [39], magnetic moments of the nucleon; model 2 (dashed curve)
which would slightly modify the leading-order prediction consists of pole terms without the anomalous magnetic
to + = (2.96 0.33) 104 fm3 instead of + = moments but including contributions from the resonances
(2.7 0.4) 104 fm3 used in [36,37]. Accordingly, the dif- (1232), P11 (1440), D13 (1520) and S11 (1535). The dotted
ference ( )+ of (12) would increase to 4.9 104 fm3 curve is a t to the experimental data.
instead of 4.4 104 fm3 , whereas the sum would remain
the same as in eq. (11).
As there is no stable pion target, empirical informa- Table 1. Previous experimental data on the charged pion po-
tion about the pion polarizabilities is not easy to obtain. larizability + .
For that purpose, one has to consider reactions which 4
contain the Compton scattering amplitude as a building Reaction Experiment + [10 fm3 ]
block, such as, e.g., the Primako eect in high-energy Z Z Serpukhov [41] 6.8 1.4 1.2
p + n Lebedev Phys. Inst. [42] 20 12
2
References [36, 37] use (l6 l5 ) = 2.7 0.4 instead of + PLUTO [44] 19.1 4.8 5.7
2.64 0.72 which was obtained in ref. [2] from = 0.44 0.12. DM 1 [45] 17.2 4.6
Correspondingly, this also generates a smaller error in the DM 2 [46] 26.3 7.4
O(q 4 ) prediction + = (2.7 0.4) 10 4 fm3 instead of MARK II [47] 2.2 1.6
(2.62 0.71) 10 4 fm3 .
d/ds1dt (nb/ ) S. Scherer: Chiral perturbation theory 63
(3) 1 0.5
Le = b23 5 [D , f ] + (25)
2(4F )2
y (fm)
0
with
-0.5
f = 2( a a ) + 2i ([v , a ] [v , a ])
i -1
+ [ , v v ] + .
F
-1.5
The Lagrangian of eq. (25) is of a non-minimal type and
-1.5 -1 -0.5 0 0.5 1 1.5
the three terms contribute to the axial-vector matrix ele- x (fm)
ment, the JA Green function and pion electroproduction
relevant to the Adler-Gilman relation. As a result it was Fig. 8. Scaled electric polarization r 3 i1 [10 3 fm3 ] [91]. The
conrmed that threshold pion electroproduction is indeed applied electric eld points in the x-direction.
a tool to obtain information on the axial form factor of
the nucleon (see [80] for details). linear sigma model, the GPs are essentially due to pionic
degrees of freedom. Due to the small pion mass the ef-
fect in the spatial distributions extends to larger distances
3.4 Virtual Compton scattering and generalized
(see also g. 9). On the other hand, the constituent quark
polarizabilities
model and other phenomenological models involving Gau
As a second example, let us discuss the application of or dipole form factors typically show a faster decrease in
HBChPT to the calculation of the so-called generalized the range Q2 < 1 GeV2 .
polarizabilities [81,82]. The virtual Compton scattering A covariant denition of the spin-averaged dipole po-
(VCS) amplitude TVCS is accessible in the reaction e p larizabilities has been proposed in ref. [55]. It was shown
e p. Model-independent predictions, based on Lorentz that three generalized dipole polarizabilities are needed to
invariance, gauge invariance, crossing symmetry, and the reconstruct spatial distributions. For example, if the nu-
discrete symmetries, have been derived in ref. [83]. Up cleon is exposed to a static and uniform external electric
to and including terms of second order in the momenta eld E, an electric polarization P is generated which is
|q | and |q | of the virtual initial and real nal photons, related to the density of the induced electric dipole mo-
the amplitude is completely specied in terms of quan- ments,
tities which can be obtained from elastic electron-proton Pi (r) = 4ij (r) Ej . (26)
scattering and real Compton scattering, namely mN , , The tensor ij (r), i.e. the density of the full electric po-
2
GE , GM , rE , p and p . The generalized polarizabilities larizability of the system, can be expressed as [55]
(GPs) of ref. [82] result from an analysis of the resid-
ual piece in terms of electromagnetic multipoles. A re- ij (r) = L (r)ri rj + T (r)(ij ri rj )
striction to the lowest-order, i.e. linear terms in leads 3ri rj ij
to only electric and magnetic dipole radiation in the - + [L (r ) T (r )] r2 dr ,
r3 r
nal state. Parity and angular-momentum selection rules,
charge-conjugation symmetry, and particle crossing gen- where L (r) and T (r) are Fourier transforms of the gen-
erate six independent GPs [82,84,85]. eralized longitudinal and transverse electric polarizabili-
The rst results for the two structure functions PLL ties L (q) and T (q), respectively. In particular, it is im-
PT T / and PLT at Q2 = 0.33 GeV2 were obtained from portant to realize that both longitudinal and transverse
a dedicated VCS experiment at MAMI [86]. Results at polarizabilities are needed to fully recover the electric po-
higher four-momentum transfer squared Q2 = 0.92 and larization P. Figure 8 shows the induced polarization in-
Q2 = 1.76 GeV2 have been reported in ref. [87]. Additional side a proton as calculated in the framework of HBChPT
data are expected from MIT/Bates for Q2 = 0.05 GeV2 at O(q 3 ) [91] and clearly shows that the polarization, in
aiming at an extraction of the magnetic polarizability. general, does not point into the direction of the applied
Moreover, data in the resonance region have been taken electric eld.
at JLab for Q2 = 1 GeV2 [88] which have been analyzed Similar considerations apply to an external magnetic
in the framework of the dispersion relation formalism of eld. Since the magnetic induction is always transverse
ref. [89,90]. Table 2 shows the experimental results of [86] (i.e., B = 0), it is su cient to consider ij (r) =
in combination with various model calculations. Clearly, (r)ij [55]. The induced magnetization M is given in
the experimental precision of [86] already allows for a crit- terms of the density of the magnetic polarizability as
ical test of the dierent models. Within ChPT and the M(r) = 4(r)B (see g. 9).
S. Scherer: Chiral perturbation theory 67
Table 2. Experimental results and theoretical predictions for the structure functions PLL PT T / and PLT at Q2 = 0.33 GeV2
and = 0.62. makes use of symmetry under particle crossing and charge conjugation which is not a symmetry of the
nonrelativistic quark model.
PLL PT T / [GeV 2
] PLT [GeV 2
]
Experiment [86] 23.7 2.2stat. 4.3syst. 0.6syst.norm. 5.0 0.8stat. 1.4syst. 1.1syst.norm.
Linear sigma model [92] 11.5 0.0
Eective Lagrangian model [93] 5.9 1.9
HBChPT [94] 26.0 5.3
Nonrelativistic quark model [95] 19.2|14.9 3.2| 4.5
0 H F (n, ) + n3 G(n, ),
2
3.6 Applications
References
1. S. Weinberg, Physica A 96, 327 (1979).
2. J. Gasser, H. Leutwyler, Ann. Phys. (N.Y.) 158, 142
(1984).
3. J. Gasser, H. Leutwyler, Nucl. Phys. B 250, 465 (1985).
4. J. Gasser, M.E. Sainio, A. Svarc, Nucl. Phys. B 307, 779
[ ] [ ] (1988).
5. G. Ecker, arXiv:hep-ph/0507056.
6. S. Scherer, in Advances in Nuclear Physics, edited
by J.W. Negele, E.W. Vogt, Vol. 27 (Kluwer Aca-
demic/Plenum, New York, 2003) pp. 277-538.
7. S. Scherer, M.R. Schindler, arXiv:hep-ph/0505265.
8. E. Jenkins, A.V. Manohar, Phys. Lett. B 255, 558 (1991).
9. V. Bernard, N. Kaiser, J. Kambor, U.-G. Mei ner, Nucl.
/
/
38. J. Bijnens, G. Colangelo, G. Ecker, J. Gasser, M.E. 73. D. Drechsel, L. Tiator, J. Phys. G 18, 449 (1992).
Sainio, Phys. Lett. B 374, 210 (1996). 74. S.L. Adler, F.J. Gilman, Phys. Rev. 152, 1460 (1966).
39. J. Bijnens, P. Talavera, Nucl. Phys. B 489, 387 (1997). 75. V. Bernard, N. Kaiser, U.G. Meissner, Phys. Rev. Lett.
40. C.Q. Geng, I.L. Ho, T.H. Wu, Nucl. Phys. B 684, 281 69, 1877 (1992).
(2004). 76. S. Scherer, J.H. Koch, Nucl. Phys. A 534, 461 (1991).
41. Y.M. Antipov et al., Phys. Lett. B 121, 445 (1983). 77. A1 Collaboration (A. Liesenfeld et al.), Phys. Lett. B
42. T.A. Aibergenov et al., Czech. J. Phys. B 36, 948 (1986). 468, 20 (1999).
43. J. Ahrens et al., Eur. Phys. J. A 23, 113 (2005). 78. D. Baumann, PhD Thesis, Johannes Gutenberg-
44. PLUTO Collaboration (C. Berger et al.), Z. Phys. C 26, Universitat, Mainz (2004).
199 (1984). 79. H. Haberzettl, Phys. Rev. Lett. 85, 3576 (2000).
45. DM1 Collaboration (A. Courau et al.), Nucl. Phys. B 80. T. Fuchs, S. Scherer, Phys. Rev. C 68, 055501 (2003).
271, 1 (1986). 81. H. Arenhovel, D. Drechsel, Nucl. Phys. A 233, 153
46. DM2 Collaboration (Z. Ajaltoni et al.), in Proceedings of (1974).
the VII International Workshop on Photon-Photon Col- 82. P.A.M. Guichon, G.Q. Liu, A.W. Thomas, Nucl. Phys. A
lisions, Paris, 1-5 April 1986, edited by A. Courau, P. 591, 606 (1995).
Kessler (World Scientic, Singapore, 1986). 83. S. Scherer, A.Y. Korchin, J.H. Koch, Phys. Rev. C 54,
47. MARK II Collaboration (J. Boyer et al.), Phys. Rev. D 904 (1996).
42, 1350 (1990). 84. D. Drechsel, G. Knochlein, A. Metz, S. Scherer, Phys.
48. D. Drechsel, L.V. Filkov, Z. Phys. A 349, 177 (1994). Rev. C 55, 424 (1997).
49. G.F. Chew, F.E. Low, Phys. Rev. 113, 1640 (1959). 85. D. Drechsel, G. Knochlein, A.Y. Korchin, A. Metz, S.
50. C. Unkmeir, PhD Thesis, Johannes Gutenberg- Scherer, Phys. Rev. C 57, 941 (1998).
Universitat, Mainz (2000). 86. VCS Collaboration (J. Roche et al.), Phys. Rev. Lett. 85,
51. L.V. Filkov, V.L. Kashevarov, Eur. Phys. J. A 5, 285 708 (2000).
(1999). 87. Jeerson Lab Hall A Collaboration (G. Laveissiere et al.),
52. C. Hacker, N. Wies, J. Gegelia, S. Scherer, Phys. Rev. C Phys. Rev. Lett. 93, 122001 (2004).
72, 055203 (2005). 88. H. Fonvieille, Prog. Part. Nucl. Phys. 55, 198 (2005).
53. C. Unkmeir, S. Scherer, A.I. Lvov, D. Drechsel, Phys. 89. B. Pasquini, M. Gorchtein, D. Drechsel, A. Metz, M. Van-
Rev. D 61, 034002 (2000). derhaeghen, Eur. Phys. J. A 11, 185 (2001).
54. T. Fuchs, B. Pasquini, C. Unkmeir, S. Scherer, Czech. J. 90. D. Drechsel, B. Pasquini, M. Vanderhaeghen, Phys. Rept.
Phys. 52, B135 (2002). 378, 99 (2003).
55. A.I. Lvov, S. Scherer, B. Pasquini, C. Unkmeir, D. Drech- 91. A.I. Lvov, S. Scherer, in preparation.
sel, Phys. Rev. C 64, 015203 (2001). 92. A. Metz, D. Drechsel, Z. Phys. A 356, 351 (1996).
56. C. Unkmeir, A. Ocherashvili, T. Fuchs, M.A. Moinester, 93. M. Vanderhaeghen, Phys. Lett. B 368, 13 (1996).
S. Scherer, Phys. Rev. C 65, 015206 (2002). 94. T.R. Hemmert, B.R. Holstein, G. Knochlein, S. Scherer,
57. L. Ametller, J. Bijnens, A. Bramon, F. Cornet, Phys. Phys. Rev. Lett. 79, 22 (1997).
Lett. B 276, 185 (1992). 95. B. Pasquini, S. Scherer, D. Drechsel, Phys. Rev. C 63,
58. P. Ko, Phys. Lett. B 349, 555 (1995). 025205 (2001).
59. S. Bellucci, C. Bruno, Nucl. Phys. B 452, 626 (1995). 96. J. Gegelia, G.S. Japaridze, K.S. Turashvili, Theor. Math.
60. A.A. Belkov, A.V. Lanyov, S. Scherer, J. Phys. G 22, Phys. 101, 1313 (1994) (Teor. Mat. Fiz. 101, 225 (1994)).
1383 (1996). 97. M.R. Schindler, J. Gegelia, S. Scherer, Nucl. Phys. B 682,
61. M. Jetter, Nucl. Phys. B 459, 283 (1996). 367 (2004).
62. J. Bijnens, A. Fayyazuddin, J. Prades, Phys. Lett. B 379, 98. T. Fuchs, M.R. Schindler, J. Gegelia, S. Scherer, Phys.
209 (1996). Lett. B 575, 11 (2003).
63. E. Oset, J.R. Pelaez, L. Roca, Phys. Rev. D 67, 073013 99. M.R. Schindler, J. Gegelia, S. Scherer, Phys. Lett. B 586,
(2003). 258 (2004).
64. S. Prakhov et al., Phys. Rev. C 72, 025201 (2005). 100. T. Fuchs, J. Gegelia, S. Scherer, Eur. Phys. J. A 19, 35
65. J. Gasser, H. Leutwyler, Nucl. Phys. B 250, 539 (1985). (2004).
66. H. Leutwyler, Phys. Lett. B 374, 181 (1996). 101. B.C. Lehnhart, J. Gegelia, S. Scherer, J. Phys. G 31, 89
67. R. Baur, J. Kambor, D. Wyler, Nucl. Phys. B 460, 127 (2005).
(1996). 102. T. Fuchs, J. Gegelia, S. Scherer, J. Phys. G 30, 1407
68. G. Knochlein, S. Scherer, D. Drechsel, Phys. Rev. D 53, (2004).
3634 (1996). 103. M.R. Schindler, J. Gegelia, S. Scherer, Eur. Phys. J. A
69. B.M.K. Nefkens et al., Phys. Rev. C 72, 035212 (2005). 26, 1 (2005).
70. J. Wess, B. Zumino, Phys. Lett. B 37, 95 (1971). 104. B. Kubis, U.-G. Mei ner, Nucl. Phys. A 679, 698 (2001).
71. E. Witten, Nucl. Phys. B 223, 422 (1983). 105. J. Friedrich, Th. Walcher, Eur. Phys. J. A 17, 607 (2003).
72. I. Giller, A. Ocherashvili, T. Ebertshauser, M.A.
Moinester, S. Scherer, Eur. Phys. J. A 25, 229 (2005).
Eur. Phys. J. A 28, s01, 71 80 (2006)
DOI: 10.1140/epja/i2006-09-008-3 EPJ A direct
electronic only
Two-photon physics
M. Vanderhaeghena
Physics Department, The College of William & Mary, Williamsburg, VA 23187, USA and Theory Group, Jeerson Lab,
12000 Jeerson Ave, Newport News, VA 23606, USA
/
Published online: 15 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. It is reviewed how Compton scattering sum rules relate low-energy nucleon structure quantities
to the nucleon excitation spectrum. In particular, the GDH sum rule and recently proposed extensions of it
will be discussed. These extensions are sometimes more calculationally robust, which may be an advantage
when estimating the chiral extrapolations of lattice QCD results, such as for anomalous magnetic moments.
Subsequently, new developments in our description of the nucleon excitation spectrum will be discussed,
in particular a recently developed chiral eective eld theory framework for the (1232)-resonance region.
Within this framework, we discuss results on N and masses, the N transition and the magnetic
dipole moment.
PACS. 25.20.Dc Photon absorption and scattering 12.39.Fe Chiral Lagrangians 13.40.Gp Electromag-
netic form factors 13.40.Em Electric and magnetic moments
crossing symmetry of the Compton amplitude of eq. (1) and, in principle, one could continue in order to isolate
means invariance under , , which obviously higher-order moments [8].
leads to f () being an even and g() being an odd function Recently, the helicity dierence which enters the
of the energy : f () = f (), g() = g(). The two integrands of eqs. (9) and (10) has been measured. The
scalar functions f (), g() admit the following low-energy rst measurement was carried out at MAMI (Mainz) for
expansion: photon energies in the range 200 MeV < < 800 MeV [9,
10], and was extended at ELSA (Bonn) [11] for up to
e2 3 GeV. This dierence, shown in g. 1, uctuates much
f () = + (E + M ) 2 + O( 4 ), (2)
4M more strongly than the total cross section T . The thresh-
e2 2 old region is dominated by S-wave pion production, and
g() = + 0 3 + O( 5 ), (3) therefore mostly contributes to the cross section 1/2 . In
8M 2
the region of the (1232) with spin J = 3/2, both helicity
and hence, in the low-energy limit, are given in terms of cross sections contribute, but since the transition is essen-
the targets charge e, mass M , and anomalous magnetic tially M 1, we nd 3/2 /1/2 3. As seen from g. 1, 3/2
moment (a.m.m.) . The next-to-leading order terms are also dominates the proton photoabsorption cross section
given in terms of the nucleon electric (E ), magnetic (M ), in the second and third resonance regions.
and forward spin (0 ) polarizabilities.
In order to derive sum rules (SRs) for these quanti-
ties one assumes the scattering amplitude is an analytic
function of everywhere but the real axis, which allows 600
writing the real parts of the functions f () and g() as 3/2-1/2 (b)
a dispersion integral involving their corresponding imagi-
500
nary parts. The latter, on the other hand, can be related to
combinations of doubly polarized photoabsorption cross-
sections via the optical theorem, 400
Im f () = 1/2 () + 3/2 () , (4) 300
8
Im g() = 1/2 () 3/2 () , (5) 200
8
where is the doubly-polarized total cross-section of the
100
photoabsorption processes, with specifying the total he-
licity of the initial system. Averaging over the polarization
of initial particles gives the total unpolarized cross-section, 0
T = 21 (1/2 + 3/2 ).
After these steps one arrives at the results (see, e.g., [1] -100
for more details): 0 200 400 600 800 1000 1200 1400 1600 1800
(MeV)
2 T ( )
f () = f (0) + 2 2
d , (6)
2 0 2 i Fig. 1. The helicity dierence 3/2 () 1/2 () for the pro-
( ) ton. The calculations include the contribution of N interme-
g() = 2 d , (7) diate states (dashed curve) [12], N intermediate state (dotted
4 0 2 2 i
curve) [13], and the N intermediate states (dashed-dotted
with 3/2 1/2 , and where the sum rule for curve) [14]. The total sum of these contributions is shown by
the unpolarized forward amplitude f () has been once- the full curves. The MAMI data are from ref. [9, 10] and the
subtracted to guarantee convergence. These relations can ELSA data from ref. [11].
then be expanded in energy to obtain the SRs for the dif-
ferent static properties introduced in eqs. (2), (3). In this
way we obtain the Baldin SR [6,7]:
2.2 Linearized GDH sum rule
1 T ()
E + M = d, (8) Recently, it was shown that by taking derivatives of the
2 2 0 2 GDH sum rule with respect to the a.m.m. one can ob-
the GDH SR: tain a new set of sum-rule like relations with intriguing
properties [4,5].
e2 2 1 () To derive such sum rules. one begins by introducing a
= d, (9)
2M 2 0 classical (or trial ) value of the particles a.m.m., 0 .
At the Lagrangian level this amounts to the introduction
a SR for the forward spin polarizability: of a Pauli term for the spin-1/2 eld :
1 () i0
0 = 2 d, (10) LPauli = F , (11)
4 0 3 4M
M. Vanderhaeghen: Two-photon physics 73
3.1 Chiral eective eld theory in the (1232) region in this case constraints do not hold for su ciently strong
electromagnetic elds, see, e.g., [50]. In extracting the
Starting from the eective Lagrangian of chiral pertur- MDM, it is therefore assumed that the electromagnetic
bation theory (PT) with pion and nucleon elds [45], eld is weak, compared to the mass scale.
the is included explicitly in the so-called -expansion The inclusion of the -resonance introduces another
scheme [39]. In the following, the Lagrangian L(i) is or- light scale besides the pion mass in the theory, the
ganized such that superscript i stands for the power of resonance excitation energy: M MN 0.3 GeV.
electromagnetic coupling e plus the number of derivatives This energy scale is still relatively light in comparison
of pion and photon elds. Writing here only the terms in- to the chiral symmetry breaking scale SB 1 GeV.
volving the spin-3/2 isospin-3/2 eld of the -isobar Therefore, = /SB can be treated as a small param-
gives:1 eter. The question is, how to compare this parameter with
the small parameter of chiral perturbation theory (PT),
(1)
LN = (i D M ) = m /SB .
In most of the literature (see, e.g., refs. [35,36,37,38,
ihA
+ N Ta ( ) D a + H.c. 40]) they are assumed to be of comparable size, .
2f M This, however, leads to a somewhat unsatisfactory result
HA because obviously the -contributions are overestimated
T a ( ) a , (13)
2M f at lower energies and underestimated at the resonance
(2) ie( 1) energies. To estimate the -resonance contributions cor-
LN = F rectly, and depending on the energy region, one needs to
2M
3iegM count and dierently.
+ N T3 F + H.c. A relation = 2 was suggested and explored in [39],
2MN (MN + M ) and is referred to as the -expansion. The second power is
ehA indeed the closest integer power for the relation of these
N Ta A a + H.c. , (14) parameters in the real world. In refs. [44,41,42,43] this re-
2f M
(3) 3e lation was used for power-counting purposes only, and was
LN = N T 3 5 [gE ( ) not imposed in the actual evaluations of diagrams. Each
2MN (MN + M )
diagram is simply characterized by an overall -counting
igC index n, which tells us that its contribution begins at
+ ( ) F + H.c., (15)
M O( n ).
Because of the distinction of m and the counting
where MN and M are, respectively, the nucleon and - of a given diagram depends on whether the characteristic
isobar masses, N and a (a = 1, 2, 3) stand for the nu- momentum p is in the low-energy region (p m ) or in
cleon and pion elds, D is the covariant derivative en- the resonance region (p ). In the low-energy region the
suring the electromagnetic gauge-invariance, F and F index of a graph with L loops, N pion propagators, NN
are the electromagnetic eld strength and its dual, Ta are nucleon propagators, N -propagators, and Vi vertices
the isospin 1/2 to 3/2 transition matrices, and T a are of dimension i is
the generators in the isospin 3/2 representation of SU (2),
satisfying T a T a = 5/3. The coupling constants are given n=2 iVi + 4L NN 2N N 2nPT N ,
by : f = 92.4 MeV, hA 2.85 is obtained from the - i
resonance width, = 0.115 GeV, and for HA the large- (16)
Nc relation HA = (9/5)gA is adopted, with gA 1.267 where nPT is the index in PT with no s [45]. In the
the nucleon axial-coupling constant. resonance region, one distinguishes the one--reducible
Note that the electric and the Coulomb N couplings (OR) graphs [39]. Such graphs contain propagators
(gE and gC , respectively) are of one order higher than the which go as 1/(p ), and hence for p they are
magnetic (gM ) one, because of the 5 which involves the large and all need to be included. This gives an incen-
small components of the fermion elds and thus intro- tive, within the power-counting scheme, to resum con-
duces an extra power of the 3-momentum. The MDM tributions. Their resummation amounts to dressing the
is dened here in units of [e/2M ]. Higher electromagnetic propagators so that they behave as 1/(p ). The
moments are omitted, because they do not contribute at self-energy begins at order p3 and thus a dressed OR
the orders that we consider. propagator counts as 1/ 3 . If the number of such propa-
(1)
Note that L contains the free Lagrangian, which is gators in a graph is NOR , the power-counting index of
formulated in [46] such that the number of spin degrees of this graph in the resonance region is given by
freedom of the relativistic spin-3/2 eld is constrained to
the physical number: 2s + 1 = 4. The N to transition n = nPT N 2NOR , (17)
couplings in eqs. (13,14,15) are consistent with these con-
straints [47,48,49]. The coupling is more subtle since where N is the total number of -propagators.
A word on the renormalization program, as it is an
1
Here we introduce totally antisymmetric products of - indivisible part of power counting in a relativistic theory.
matrices: = 12 [ , ], = 12 { , } = i 5 . Indeed, without some kind of renormalization the loop
M. Vanderhaeghen: Two-photon physics 75
M (GeV)
counting index of the graph. Also, contributions of heavy
scales, such as baryon masses, may appear as M N . The 1.7
renormalization of the loop graphs can and should be per- 1.6
formed so as to absorb these large contributions into the 1.5
available low-energy constants, thus bringing the result in 1.4
accordance with power counting [51]. 1.3
To give an example, consider the one-N -loop con- 1.2
tribution to the nucleon mass. For the N N vertex, 1.1
the power counting tells us that this contribution be- 1 N
gins at O(m3 ). An explicit calculation, however, will show 0.9
(e.g., [45]) that the loop produces O(m0 ) and O(m2 )
0.8
terms, both of which are (innitely) large. This is not a 0 0.1 0.2 0.3 0.4 0.5
violation of power counting, because there are two low- 2
m (GeV ) 2
energy constants: the nucleon mass in the chiral limit,
M (0) , and c1N , which enter at order O(m0 ) and O(m2 ),
respectively, and renormalize away the large contributions Fig. 3. Pion-mass dependence of the nucleon and (1232)
coming from the loop. The renormalized relativistic result, masses. The curves are two-parameter expressions for the N
up to and including O(m3 ), can be written as [44]: loop contributions to MN and M as calculated in ref. [44]
(see text for the values of the low-energy constants). The red
(0) squares are lattice results from the MILC Collaboration [52].
MN = MN 4 c1N m2 (18)
5/2
The stars represent the physical mass values.
2
3 gA m2 m
2
m3 4 1 2 arccos
(8f ) 4MN 2MN
3
17m m
+
16MN 2MN (a) (b)
2 4
m m m m
+ 30 10 + ln ,
8MN MN MN MN
(c) (d) (e) (f)
and one can easily verify that the loop contribution begins Fig. 4. Diagrams for the eN eN reaction at NLO in the
at O(m3 ) in agreement with power counting. -expansion. Double lines represent the propagators.
Likewise, the mass has also been calculated in rela-
tivistic EFT see ref. [44] for details.
The m dependence of the nucleon and -resonance g. 4(c f). The hadronic part of graph (a) begins at O( 0 )
masses are compared with lattice results in g. 3. One of which here is the leading order. The Born graphs (b) con-
the two parameters in eq. (18) is constrained by the physi- tribute at O(). The one-loop vertex corrections of g. 4(e)
cal nucleon mass value at m = 0.139 GeV, while the other and (f) to the N -transition form factors have been eval-
parameter is t to the lattice data shown in the gure. uated in two independent ways in refs. [41,42], to which
This yields : MN = 0.883 GeV and c1N = 0.87 GeV1 .
(0) we refer for details. At NLO there are also vertex cor-
As is seen from the gure, with this two-parameter form rections of the type (e) and (f) with nucleon propagators
for MN , a good description of lattice results is obtained up in the loop replaced by the -propagators. However, after
to m2 0.5 GeV2 . Analogously to the nucleon case, one the appropriate renormalizations and Q2 SB , these
low-energy constant for the is xed from the physical graphs start to contribute at next-next-to-leading order.
value of the mass, while the second parameter is t to The vector-meson diagram, g. 4(d), contributes to NLO
(0) for Q2 SB . It was included eectively in refs. [41,42]
the lattice data shown in g. 3, yielding : M = 1.20 GeV by giving the gM -term a dipole Q2 -dependence, in anal-
and c1 = 0.40 GeV1 . As well as for the nucleon, this ogy to how it is usually done for the nucleon isovector
two-parameter form for M yields a fairly good descrip- form factor.
tion of the lattice results up to m2 0.5 GeV2 . The resonant pion photoproduction multipoles are
used to determine the two low-energy constants: gM and
gE , the strength of the M 1 and E2 N transitions.
3.2 N transition In g. 5, we show the result of the EFT calculations
(3/2)
The N transition is usually studied through the pion for the pion photoproduction resonant multipoles M1+
(3/2)
electroproduction process. The pion electroproduction and E1+ , around the resonance position, as function of
amplitude to NLO in the expansion, in the resonance the total c.m. energy W of the N system. These two
region, is given by graphs in g. 4(a) and (b), where the multipoles are well established by the MAID [12] and
shaded blobs in graph (a) include corrections depicted in SAID [53] partial-wave solutions which allow us to t
76 The European Physical Journal A
Re M1+3/2 (10-3/m)
30 4
40
T + L (b/sr)
20
2
LT (b/sr)
10
0 0
20
-10
-2
-20
-4
Im M1+3/2 (10-3/m)
0
40
10 6
30
LT (b/sr)
TT (b/sr)
0 4
20
-10 2
10
0 -20 0
1.18 1.2 1.22 1.24 1.26 1.28
W (GeV) -30 -2
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
cos cos
1
Re E1+3/2 (10-3/m)
0.5
obtains the dotted curves in g. 5. These non-resonant
0
contributions are purely real at this order and do not af-
-0.5
fect the imaginary part of the multipoles. One sees that
-1 the resulting calculation is awed because the real parts
-1.5 of the resonant multipoles now fail to cross zero at the res-
-2 onance position and hence unitarity, in the sense of Wat-
-2.5 sons theorem [55], is violated. The complete NLO calcu-
1.18 1.2 1.22 1.24 1.26 1.28
lation, shown by the solid curves in g. 5, includes in addi-
W (GeV) tion the vertex corrections, g. 4(e, f), which restore uni-
tarity exactly. Watsons theorem is satised exactly by the
(3/2) (3/2)
Fig. 5. Multipole amplitudes M1+ (top panels) and E1+ NLO, up to-one-loop amplitude given the graphs in g. 4.
(bottom panels) for pion photoproduction as function of the in- Figure 6 shows the NLO results for dierent vir-
variant mass W of the N system. Dashed curves: contribu- tual photon absorption cross sections (for denitions,
tion without the N -vertex corrections, (i.e., g. 4(a) with- see ref. [42]) at the resonance position, and for Q2
out g. 4(e, f)). Dotted curves: adding the Born contributions, 0.127 GeV2 , where recent precision data are available. Be-
g. 4(b), to the dashed curves. Solid curves: complete NLO
sides the low-energy constants gM and gE , which were
calculation, includes all graphs from g. 4. In all curves the
xed from the resonant multipoles in g. 5, the only other
low-energy parameters are chosen as : gM = 2.9, gE = 1.0.
The data point are from the SAID analysis (FA04K) [53] (red
low-energy constant from eq. (15) entering the NLO elec-
circles), and from the MAID 2003 analysis [12] (blue squares). troproduction calculation is gC . The main sensitivity on
gC enters in LT . A best description of the LT data in
g. 6 is obtained by choosing gC = 2.36.
the two low-energy constants of the chiral Lagrangian of The theoretical uncertainty due to the neglect of
eqs. (14,15) as : gM = 2.9, gE = 1.0. As is seen from higher-order eects was estimated in ref. [42]. We know
g. 5, with these values the NLO results (solid lines) give that they must be suppressed by at least one power of
a good description of the energy dependence of the reso- (= /SB ) as compared to the NLO and two powers of
nant multipoles in a window of 100 MeV around the - as compared to the LO contributions. These error esti-
resonance position. Also, these values yield REM = 2.3 mates are shown by the bands in g. 6. One sees that the
%, in a nice agreement with experiment [16]. NLO EFT calculation, within its accuracy, is consistent
The dashed curves in g. 5 show the contribution of with the experimental data for these observables.
the -resonant diagram of g. 4(a) without the NLO ver- Figure 7 shows the Q2 dependence of the ratios REM
tex corrections g. 4(e, f). For the M1+ multipole this is and RSM . Having xed the low energy constants gM , gE
the LO contribution. For the E1+ multipole the LO con- and gC , this Q2 dependence follows as a prediction. The
tribution is absent (the gE coupling is of one order higher theoretical uncertainty here (shown by the error bands)
than gM ). Hence, the dashed curve represents a partial was also estimated in ref. [42] over the range of Q2 from 0
NLO contribution to E1+ therein. Upon adding the non- to 0.2 GeV2 . One sees that the NLO calculations are con-
resonant Born graphs, g. 4(b), to the dashed curves one sistent with the experimental data for both of the ratios.
M. Vanderhaeghen: Two-photon physics 77
4 Conclusions
2 Elab = 400 MeV
0.1
present lattice QCD results obtained at unphysical values
of m to the experiment. This was demonstrated here ex-
0 E lab = 400 MeV plicitely for the N and masses, the N transition and
c.m. = 90o the magnetic dipole moment. As the next-generation
-0.1
0 20 40 60 80 100 120 140 160 lattice calculations of these quantities are on the way [69],
E c.m. (MeV)
such a EFT framework will, hopefully, complement these
eorts.
Fig. 11. The outgoing photon energy dependence of the
p 0 p observables for dierent values of + (in units
e/2M ). Top panel: the ratio of p 0 p to p 0 p cross- I am grateful to my colleagues in Mainz for the unique culture
sections eq. (19). Data points are from [62]. Middle panel: the of cross-fertilization between experiment and theory. On the
linear-polarization photon asymmetry of the p 0 p cross- subject of two-photon physics, I like to thank in particular
sections dierential w.r.t. the outgoing photon energy and pion Dieter Drechsel and Barbara Pasquini, for the many collabo-
c.m. angle. The data point at E = 0 corresponds with the rations. I also like to acknowledge Vladimir Pascalutsa for very
p 0 p photon asymmetry from [16]. Lower panel: the fruitful recent collaborations on the EFT in the -resonance
circular-polarization photon asymmetry (as dened in [64]), region. This work is supported in part by DOE grant no. DE-
where the outgoing photon angles have been integrated over FG02-04ER41302 and contract DE-AC05-84ER-40150 under
the indicated range. which SURA operates the Jeerson Laboratory.
References
process. Mechanisms involving -photoproduction Born
1. D. Drechsel, B. Pasquini, M. Vanderhaeghen, Phys. Rep.
terms followed by N rescattering have been considered
378, 99 (2003).
in model calculations [63,64]. In the -counting they start
2. S.B. Gerasimov, Sov. J. Nucl. Phys. 2, 430 (1966) (Yad.
contributing at next-next-to-leading order and therefore Fiz. 2, 598 (1966)); S.D. Drell, A.C. Hearn, Phys. Rev.
will provide the main source of corrections to the present Lett. 16, 908 (1966).
NLO results. 3. D. Drechsel, L. Tiator, Annu. Rev. Nucl. Part. Sci. 54, 69
(2004).
The asymmetry for circularly polarized photons, circ ,
4. V. Pascalutsa, B.R. Holstein, M. Vanderhaeghen, Phys.
(which is exactly zero for a two-body process due to re- Lett. B 600, 239 (2004).
ection symmetry w.r.t. the reaction plane) has been pro- 5. B.R. Holstein, V. Pascalutsa, M. Vanderhaeghen, Phys.
posed [64] as a unique observable to enhance the sensi- Rev. D 72, 094014 (2005).
tivity to . Indeed, in the soft-photon limit, where the 6. A.M. Baldin, Nucl. Phys. 18, 310 (1960).
p 0 p process reduces to a two-body process, circ 7. L.I. Lapidus, Sov. Phys. JETP 16, 964 (1963).
is exactly zero. Therefore, its value at higher outgoing pho- 8. B.R. Holstein, D. Drechsel, B. Pasquini, M. Vander-
ton energies is directly proportional to . One sees from haeghen, Phys. Rev. C 61, 034316 (2000).
g. 11 (lower panel) that our EFT calculation supports 9. GDH and A2 Collaborations (J. Ahrens et al.), Phys. Rev.
this observation, and shows sizeably dierent asymmetries Lett. 84, 5950 (2000).
for dierent values of . A combined t of all three ob- 10. GDH and A2 Collaborations (J. Ahrens et al.), Phys. Rev.
servables shown in g. 11 will therefore allow for a very Lett. 87, 022003 (2001).
stringent test of the EFT calculation, which can then be 11. GDH Collaboration (H. Dutz et al.), Phys. Rev. Lett. 91,
used to extract the + MDM. 192001 (2003).
80 The European Physical Journal A
12. D. Drechsel, O. Hanstein, S.S. Kamalov, L. Tiator, Nucl. 35. E. Jenkins, A.V. Manohar, Phys. Lett. B 255, 558 (1991);
Phys. A 645, 145 (1999). 259, 353 (1991).
13. D. Drechsel, S.S. Kamalov, L. Tiator, Phys. Rev. D 63, 36. T. Hemmert, B.R. Holstein, J. Kambor, Phys. Lett. B 395,
114010 (2001). 89 (1997); G. Gellas et al., Phys. Rev. D 60, 054022 (1999).
14. H. Holvoet, PhD Thesis, University Gent (2001). 37. H.-B. Tang, P. Ellis, Phys. Lett. B 387, 9 (1996).
15. D.A. Dicus, R. Vega, Phys. Lett. B 501, 44 (2001). 38. N. Fettes, U.G. Meissner, Nucl. Phys. A 679, 629 (2001).
16. R. Beck et al., Phys. Rev. Lett. 78, 606 (1997); Phys. Rev. 39. V. Pascalutsa, D.R. Phillips, Phys. Rev. C 67, 055202
C 61, 035204 (2000). (2003); 68, 055205 (2003).
17. G. Blanpied et al., Phys. Rev. Lett. 79, 4337 (1997). 40. C. Hacker, N. Wies, J. Gegelia, S. Scherer, Phys. Rev. C
18. C. Mertz et al., Phys. Rev. Lett. 86, 2963 (2001); N.F. 72, 055203 (2005).
Sparveris et al., Phys. Rev. Lett. 94, 022003 (2005). 41. V. Pascalutsa, M. Vanderhaeghen, Phys. Rev. Lett. 95,
19. V.V. Frolov et al., Phys. Rev. Lett. 82, 45 (1999); K. Joo 232001 (2005).
et al., Phys. Rev. Lett. 88, 122001 (2002). 42. V. Pascalutsa, M. Vanderhaeghen, arXiv:hep-ph/0512244.
20. S.L. Glashow, Physica A 96, 27 (1979); N. Isgur, G. Karl, 43. V. Pascalutsa, M. Vanderhaeghen, Phys. Rev. Lett. 94,
R. Koniuk, Phys. Rev. D 25, 2394 (1982); S. Capstick, 102003 (2005).
G. Karl, Phys. Rev. D 41, 2767 (1990); S. Capstick, B.D. 44. V. Pascalutsa, M. Vanderhaeghen, arXiv:hep-ph/0511261.
Kiester, Phys. Rev. D 51, 3598 (1995); G.A. Miller, Phys. 45. J. Gasser, M.E. Sainio, A. Svarc, Nucl. Phys. B 307, 779
Rev. C 68, 022201(R) (2003); A.M. Bernstein, Eur. Phys. (1988).
J. A 17, 349 (2003). 46. W. Rarita, J.S. Schwinger, Phys. Rev. 60, 61 (1941).
21. M.G. Olsson, E.T. Osypowski, Nucl. Phys. B 87, 399 47. V. Pascalutsa, Phys. Rev. D 58, 096002 (1998).
(1974); Phys. Rev. D 17, 174 (1978); R.M. Davidson, 48. V. Pascalutsa, R.G.E. Timmermans, Phys. Rev. C 60,
N.C. Mukhopadhyay, R.S. Wittman, Phys. Rev. D 43, 71 042201(R) (1999).
(1991). 49. V. Pascalutsa, Phys. Lett. B 503, 85 (2001).
22. H. Garcilazo, E. Moya de Guerra, Nucl. Phys. A 562, 521 50. S. Deser, V. Pascalutsa, A. Waldron, Phys. Rev. D 62,
(1993); C. Fernandez-Ramirez, E. Moya de Guerra, J.M. 105031 (2000).
Udias, arXiv:nucl-th/0509020. 51. J. Gegelia, G. Japaridze, Phys. Rev. D 60, 114038 (1999);
23. M. Vanderhaeghen, K. Heyde, J. Ryckebusch, M. Waro- J. Gegelia, G. Japaridze, X.Q. Wang, J. Phys. G 29, 2303
quier, Nucl. Phys. A 595, 219 (1995). (2003).
24. V. Pascalutsa, O. Scholten, Nucl. Phys. A 591, 658 (1995); 52. C.W. Bernard et al., Phys. Rev. D 64, 054506 (2001).
O. Scholten, A.Y. Korchin, V. Pascalutsa, D. Van Neck, 53. R.A. Arndt, W.J. Briscoe, I.I. Strakovsky, R.L. Workman,
Phys. Lett. B 384, 13 (1996); A.Y. Korchin, O. Scholten, Phys. Rev. C 66, 055213 (2002).
R.G.E. Timmermans, Phys. Lett. B 438, 1 (1998). 54. C. Kunz et al., Phys. Lett. B 564, 21 (2003).
25. T. Feuster, U. Mosel, Phys. Rev. C 59, 460 (1999); G. 55. K.M. Watson, Phys. Rev. 95, 228 (1954).
Penner, U. Mosel, Phys. Rev. C 66, 055211 (2002); H. 56. T. Pospischil et al., Phys. Rev. Lett. 86, 2959 (2001).
Lenske, V. Shklyar, U. Mosel, arXiv:nucl-th/0512044. 57. D. Elsner et al., arXiv:nucl-ex/0507014.
26. I.G. Aznauryan, Phys. Rev. C 68, 065204 (2003). 58. C. Alexandrou et al., Phys. Rev. Lett. 94, 021601 (2005).
27. S. Nozawa, B. Blankleider, T.-S.H. Lee, Nucl. Phys. A 513, 59. B.M.K. Nefkens et al., Phys. Rev. D 18, 3911 (1978).
459 (1990); T.-S.H. Lee, B.C. Pearce, Nucl. Phys. A 530, 60. A. Bosshard et al., Phys. Rev. D 44, 1962 (1991).
532 (1991); T. Sato, T.-S.H. Lee, Phys. Rev. C 54, 2660 61. D. Drechsel, M. Vanderhaeghen, M.M. Giannini, E. San-
(1996); 63, 055201 (2001). topinto, Phys. Lett. B 484, 236 (2000).
28. Y. Surya, F. Gross, Phys. Rev. C 53, 2422 (1996); 47, 703 62. M. Kotulla et al., Phys. Rev. Lett. 89, 272001 (2002).
(1993). 63. D. Drechsel, M. Vanderhaeghen, Phys. Rev. C 64, 065202
29. S.S. Kamalov, S.N. Yang, Phys. Rev. Lett. 83, 4494 (1999); (2001).
S.S. Kamalov, G.Y. Chen, S.N. Yang, D. Drechsel, L. Tia- 64. W.T. Chiang, M. Vanderhaeghen, S.N. Yang, D. Drechsel,
tor, Phys. Lett. B 522, 27 (2001); L. Tiator, D. Drechsel, Phys. Rev. C 71, 015204 (2005).
S.S. Kamalov, S.N. Yang, Eur. Phys. J. A 17, 357 (2003). 65. R. Beck, B. Nefkens, spokespersons Crystal Ball at MAMI
30. M.G. Fuda, H. Alharbi, Phys. Rev. C 68, 064002 (2003). experiment.
31. V. Pascalutsa, J.A. Tjon, Phys. Lett. B 435, 245 (1998); 66. Particle Data Group (K. Hagiwara et al.), Phys. Rev. D
Phys. Rev. C 61, 054003 (2000); 70, 035209 (2004). 66, 010001 (2002).
32. G. Caia, V. Pascalutsa, J.A. Tjon, L.E. Wright, Phys. Rev. 67. D.B. Leinweber, T. Draper, R.M. Woloshyn, Phys. Rev.
C 70, 032201(R) (2004); G. Caia, L.E. Wright, V. Pasca- D 46, 3067 (1992); I.C. Cloet, D.B. Leinweber, A.W.
lutsa, Phys. Rev. C 72, 035203 (2005). Thomas, Phys. Lett. B 563, 157 (2003).
33. S. Weinberg, Physica A 96, 327 (1979). 68. F.X. Lee, R. Kelly, L. Zhou, W. Wilcox, arXiv:hep-
34. J. Gasser, H. Leutwyler, Ann. Phys. (N.Y.) 158, 142 lat/0410037.
(1984). 69. C. Alexandrou et al., arXiv:hep-lat/0509140.
Eur. Phys. J. A 28, s01, 81 90 (2006)
DOI: 10.1140/epja/i2006-09-009-2 EPJ A direct
electronic only
/
Published online: 12 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Elastic form factors are of fundamental importance for the understanding of microscopic spatial
structures. In case of the proton and the neutron, charge and magnetic form factors can be studied in
elastic electron scattering. Techniques to accelerate polarised continuous electron beams, the availability of
polarised targets as well as modern concepts and instrumentation for coincidence experiments and recoil
polarimetry had an enormous impact on these measurements. The developments and experiments at the
Mainz Microtron MAMI will be discussed in a general context.
PACS. 13.40.Gp Electromagnetic form factors 13.85.Dz Elastic scattering 13.88.+e Polarization in
interactions and scattering 25.30.Bf Elastic electron scattering
y
reaction plane
x pe pn
z
q R
pe pp
d(e,en), Bates 93
Rn, NIKHEF 94
Rn, ELSA 95 nq
Rn, MAMI 98, 02
3
He(e,e), Bates 94
3
He(e,e), Jlab 00
electron scattering plane
3
He(e,e), Jlab 03 Fig. 3. Reference frame and kinematics of the d(e, e n) reac-
tion.
2 2
Q /(GeV/c)
where the contribution of GE,p to the unpolarised cross
Fig. 2. The neutron magnetic form factor in units of n GD section is kinematically suppressed. The increased sensi-
as function of Q2 measured in coincidence and polarisation tivity of double polarisation observables to GE,n and GE,p
experiments [8, 11, 12, 13, 14, 19, 20, 21]. at high Q2 has already been pointed out more than 40
years ago [23,24,25].
For the ideal case of scattering longitudinally polarised
electron o free nucleons, N (
are summarised in g. 2 in units of the empirical dipole
e , e N ), the components of
expression (eq. (3)).
the recoil polarisation are given by
An alternative method to determine GM,n is provided
by inclusive scattering of polarised electrons from po-
2 (1 ) GE GM
larised 3 He in quasi-elastic kinematics [19,20,21]. Results Px = Pe , (6)
obtained with this technique at Bates and Jeerson Lab G2E + G2M
are included in g. 2 as well as absolute d(e, e n) cross Py = 0 , (7)
section measurements from Bates [8].
1 GM
2 2
Recently, new measurements of Rd at Q2 values up to Pz = P e . (8)
5 (GeV/c)2 have been completed at Jeerson Lab [22]. G2E + G2M
The large solid angle covered by the CLAS spectrome-
They are equivalent to cross section asymmetries with re-
ter allows to perform the e ciency determination simul-
spect to the beam helicity for the dierent nucleon spin
taneously with the Rd measurement by tagging neutrons
orientations in the scattering from polarised targets:
in the p(e, e + )n reaction where both, scattered electron
and + are detected. Preliminary results show, that GM,n Ax = Px ; Ay = 0; Az = P z . (9)
follows the dipole approximation up to Q2 = 5 (GeV/c)2
within 10%. The x and z direction are dened by the electron scatter-
ing plane with z given by the momentum transfer q (see
g. 3).
3 Double-polarisation observables In Px and Ax both form factors enter linearly which
increases the sensitivity compared to the unpolarised cross
Experiments using polarised electron beams in combina- section, if G2E G2M .
tion with polarised nucleons either in the initial or nal In case of the neutron, the free e-n scattering has again
state oer possibilities to measure interferences between to be approximated by the quasi-free scattering o light
longitudinal and transverse amplitudes which do not ap- nuclei (2 H, 3 He) and one has to pay attention to nuclear
pear in the unpolarised cross section. This is particularly binding and rescattering eects. In leading order, spectral
interesting in cases where one part is completely dom- functions cancel in the polarisation and asymmetry com-
inating and unpolarised cross section measurements are ponents being ratios of cross sections. However, higher
not su cient to separate additional small amplitudes. Fur- order eects like FSI and MEC as well as inuences of
thermore, most polarisation observables are insensitive to Fermi-motion on the projections of polarisation compo-
absolute luminosities and other experimental calibration nents have to be taken into account.
factors. Polarised 3 He can be used as an eective polarised neu-
In electron-neutron scattering for example, the small- tron target because in its ground state the two protons are
ness of the electric form factor GE,n compared to the dominantly in the s-state with the spins coupled to zero.
dominant magnetic form factor makes a reliable Rosen- Thus the spin of the 3 He is predominantly carried by the
bluth separation impossible. As mentioned above, the sit- neutron. Additional d-wave components, meson exchange
uation is similar for protons at high momentum transfer currents and nal state interactions have recently been
84 The European Physical Journal A
0.2 o
R=180 4.1 Polarised electrons at MAMI
Px in percent
1.2 Pospischil et al., MAMI(2001)
Milbrath et al., Bates (1999)
Jones et al., JLab (2000)
40 Simon et al., Mainz (1980)
G Ep / (G Mp / p)
1.1
30
1
20
10
eff
0.9
Px
0
0 2 4 6 8 10 12 0.35 0.4 0.45 0.5
c.m.
np Q 2 (GeV 2/c 2)
Fig. 5. Transverse proton polarisation Px measured in Fig. 6. The ratio p GE,p /GM,p from polarisation transfer
p(
e ,e p ) (square) and d( e ,e
p ) (circles) at Q2 = measurements at low Q2 [36, 37, 38]. The dotted line corre-
0.3 (GeV/c)2 [35]. In case of the deuteron target the observed sponds to the exact scaling behaviour of eq. (4), the solid line
dependence on the angle np of the proton-neutron relative is a t to Rosenbluth separated data [3].
momentum is shown. The line is a calculation of Arenhovel et
al. [27, 28].
1 A( ) N + ( ) N ( + )
= , (11)
1 + A( ) N ( ) N + ( + )
a T () in % 0.12
GE,n ,
D(e,e n), Bates
A eff= 20% 0.1
,
D(e,e n), MAMI, A3
3 ,
He(e,e n), MAMI, A1
3 ,
5 0.08 He(e,e n), MAMI, A3
,
D(e,e n), NIKHEF
A eff= 8% 0.06
0
0.04
5
0 0.02
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
10
100 50 0 50 100 2 2
Q / (GeV/c)
in degree
Fig. 10. First GE,n results from double-polarisation observ-
Fig. 9. Measured azimuthal asymmetries a () for various ables. The arrows indicate the inuence of few-body eects
precession angles with two dierent cuts on the analysing mainly due to nal state interactions. The shaded area repre-
reaction leading to dierent analysing powers but leaving 0 sents the model dependence of GE,n values extracted from elas-
unchanged. tic D(e, e ) experiments [50] The dashed line is the parametri-
sation of Galster et al. [49].
G E,n r 2 (r) fm
Bates
0.015
3
MAMI He(e,e n) MAMI
0.1 D(e,en)
JLAB
0.01
NIKHEF D(e,e)D
0.08 D(e,en)
JLAB
0.005
0.06
0
0.04
0.005
0.02
0 0 1 2 3 4 5
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Q 2 / (GeV/c) 2 r/fm
Fig. 11. Present status of GE,n measurements compared to Fig. 12. Neutron charge distribution obtained from Fourier
the parametrisation of Galster et al. ([49], dashed line) and transforms of GE,n by Friedrich and Walcher (solid line [63])
Friedrich, Walcher ([63], solid line). The data are from experi- and the Galster parametrisation (dashed line).
ments using polarised Deuterium [64, 65, 59], polarised 3 He [55,
26, 58], recoil polarisation [66, 54, 60], and from an analysis of
the elastic deuteron quadrupole form factor [62]. imately 0.2 (GeV/c)2 as well as a corresponding bump in
GE,n are clearly revealed [63].
Nonrelativistically, if the Compton wavelength of a
above-mentioned older analyses of the elastic deuteron system
is negligible compared to its size C = h/M c
structure functions A [50], the model uncertainties are re- r2 , form factors can be measured over a su ciently
duced by the direct use of FQ . large range in momentum transfer in order to calculate
Taking all these novel approaches together, a consis- spatial densities by a Fourier transform without relativis-
tent picture of the charge form factor of the neutron is tic eects becoming important. This is the case for atomic
starting to arise (see g. 11). Results obtained with dif- nuclei and detailed information about nuclear charge dis-
ferent targets and in dierent reactions are in fair agree- tributions has been obtained from electron scattering [67].
ment with each other even though at low Q2 substantial Presently, also the extraction of mass or neutron densities
corrections due to rescattering are unavoidable. The data are discussed [68,69].
roughly follow a phenomenological parametrisation given Although for nucleons C 0.25 r2 , a similar in-
by Galster et al. already in 1971 [49]. However, the accu- terpretation of GE (Q2 ) and GM (Q2 ) as momentum repre-
racy is reaching a level at which deviations from such a sentations of spatial charge and magnetisation densities is
simple, smooth behavior start to become signicant [63]. still possible in the Breit frame of vanishing energy trans-
fer. Figure 12 shows the corresponding charge distribution
of the neutron as calculated by Friedrich and Walcher from
their ts.
5 Interpretation
In coordinate space, the structures observed around
Q2 = 0.25 (GeV/c)2 inuence the long distance tail
With modern experimental techniques, for the rst time (r 1.5 2 fm) of charge and magnetisation densities and
all elastic nucleon form factors, including the neutron may be interpreted as resulting from a pion cloud sur-
charge form factor, have been measured precisely over a rounding a bare nucleon. Within this picture, the data for
nite range in momentum transfer. all four form factors can be described by an intuitive phe-
Both magnetic form factors, GM,p and GM,n , fol- nomenological ansatz consisting of dipole functions for the
low the dipole approximation within 10% up to Q2 = constituent quarks together with a p-shell harmonic oscil-
5 (GeV/c)2 . The scaling relation (eq. (4)) is violated lator behaviour of the pion cloud [63].
considerably for the proton electric form factor. The al- Another method to analyse and interpret form fac-
most linear decrease of the ratio GE,p /GM,p at Q2 > tors, which does not directly refer to a particular model
1 (GeV/c)2 , as revealed by spin transfer measurements for nucleon structure, is based on dispersion relations in
at JLab, indicates that the charge density of the proton is Q2 . They provide a mathematical framework to connect
signicantly softer than its magnetisation density. experimental date in spacelike (Q2 > 0) as well as in
Friedrich and Walcher emphasised local deviations timelike (Q2 < 0) regions with spectral functions describ-
from a smooth shape in the Q2 dependence of all four ing the spectrum of virtual intermediate states, through
form factors. By tting the available data with an ansatz which a photon can couple to a nucleon [70]. Already
given by the sum of a Gaussian and two dipoles dis- early form factor data have been analysed systematically
cribing the smooth part, local minima in GE,p , GM,p and in this framework and it has been established that the
GM,n around Q2 = 0.25 (GeV/c)2 with a width of approx- spectral functions can be approximated by poles due to
M. Ostrick: Electromagnetic form factors of the nucleon 89
the existence of vector mesons and their coupling to nu- 5. C.E. Hyde-Wright, K. de Jager, Annu. Rev. Nucl. Part.
cleons. The prediction of the (770) meson was an early Sci. 54, 217 (2004).
success of this approach [71]. 6. H.F. Ehrenberg, R. Hofstadter, Phys. Rev. 110, 544
Rened analyses demonstrated the importance of non (1958).
resonant multi-pion intermediate states [72,73,74]. Two- 7. A. Lung et al., Phys. Rev. Lett. 70, 718 (1993).
and tree-pion systems are the lightest possible intermedi- 8. P. Markowitz et al., Phys. Rev. C 48, R5 (1993).
ate states. They provide a link to pion-nucleon scattering 9. W. Fabian, H. Arenhovel, Nucl. Phys. A 314, 253 (1979).
and to model-independent predictions from chiral pertur- 10. M. Schwamb, these proceedings.
bation theory. The analysis method based on dispersion 11. H. Anklin et al., Phys. Lett. B 336, 313 (1994).
12. E.E.W. Bruins et al., Phys. Rev. Lett. 75, 21 (1995).
relations as well as the inuence of recent data on the
13. H. Anklin et al., Phys. Lett. B 428, 248 (1998).
spectral functions is discussed by H.W. Hammer [75].
14. G. Kubon et al., Phys. Lett. B 524, 26 (2002).
15. K.I. Blomqvist et al., Nucl. Instrum. Methods A 403, 263
6 Conclusions (1998).
16. J. Arnold et al., Nucl. Instrum. Methods A 386, 211
In facilities like the Mainz Microtron MAMI, high inten- (1997).
sity, polarised, continuous electron beams in the energy 17. J. Jourdan, I. Sick, J. Zhao, Phys. Rev. Lett. 79, 5186
(1997).
range relevant to study phenomena at hadronic scales
18. E.E.W. Bruins et al., Phys. Rev. Lett. 79, 5187 (1997).
are available and can be combined with polarised targets
19. H. Gao et al., Phys. Rev. C 50, R546 (1994).
and sophisticated detector systems for coincidence exper- 20. W. Xu et al., Phys. Rev. Lett. 85, 2900 (2000).
iments and polarimetry. 21. W. Xu et al., Phys. Rev. C 67, 012201 (2003).
Electromagnetic form factors are signicant observ- 22. W.K. Brooks, J.D. Lachniet, Nucl. Phys. A 755, 261
ables, directly related to the spatial structure of the nu- (2005).
cleon. For the rst time, all four nucleon form factors have 23. A.I. Akhiezer et al., Sov. Phys. JETP 6, 588 (1958).
been measured with a precision su cient to identify local 24. N. Dombey, Rev. Mod. Phys. 41, 236 (1969).
structures in the Q2 dependence at a few percent level. 25. R.G. Arnold, C.E. Carlson, F. Gross, Phys. Rev. C 23, 363
In the near future, measurements of GE,p /GM,p and (1981).
GE,n will be extended to higher values of Q2 at Jeerson 26. J. Golak et al., Phys. Rev. C 63, 034006 (2001).
Lab. Below Q2 = 2 (GeV/c)2 , new 3 H e( e , e n) as well as 27. H. Arenhovel et al., Z. Phys. A 331, 123 (1988).
28. H. Arenhovel et al., Phys. Rev. C 52, 1232 (1995).
absolute p(e, e ) cross section measurements are planned
at MAMI [76,77]. 29. K. Aulenbacher et al., Nucl. Instrum. Methods A 391, 498
Besides electromagnetic form factors, a deeper under- (1997).
30. K.H. Steens et al., Nucl. Instrum. Methods A 325, 378
standing of the elastic nucleon response also includes the
(1993).
weak vector and axial-vector currents. The nucleon axial
31. V. Tioukine et al., contribution to the 8th European Par-
form factor at low Q2 has recently been measured in pion ticle Accelerator Conference (EPAC 2002), Paris, France,
electroproduction at MAMI [78] and present-day exper- 3-7 June 2002.
iments in parity-violating electron scattering provide ac- 32. P. Drescher et al., Nucl. Instrum. Methods A 381, 169
cess to the two weak vector form factors, which will allow (1996).
a avour decomposition of the charge and magnetisation 33. A. Jankowiak, these proceedings.
distributions in the nucleon [34,69]. 34. F. Maas, these proceedings.
Also beyond elastic scattering, the experimental tech- 35. D. Eyl et al., Z. Phys. A 352, 211 (1995).
niques discussed above help to fully exploit the properties 36. B.D. Milbrath et al., Phys. Rev. Lett. 80, 452 (1998).
of electromagnetic probes for studies of the much poorer 37. T. Pospischil et al., Eur. Phys. J. A 12, 125 (2001).
known structure and dynamics of resonances in inelastic 38. M.K. Jones et al., Phys. Rev. Lett. 84, 1398 (2000).
processes [79,80]. 39. O. Gayou et al., Phys. Rev. Lett. 88, 092301 (2002).
40. J. Arrington, Phys. Rev. C 69, 022201 (2004).
41. I.A. Qattan et al., Phys. Rev. Lett. 94, 142301 (2005).
I would like to thank the organisers of the symposium 20 Years 42. L.W. Mo, Y.S. Tsai, Rev. Mod. Phys. 41, 205 (1969).
of Physics at the Mainz Microtron MAMI, Hartmut Arenhovel, 43. M. Vanderhaeghen, these proceedings.
Hartmut Backe, Dieter Drechsel, Jorg Friedrich, Karl-Heinz 44. E. Fermi, L. Marshal, Phys. Rev. 72, 1139 (1947).
Kaiser and Thomas Walcher and express all the best wishes 45. S. Kopecki et al., Phys. Rev. C 56, 2229 (1997).
for the future. 46. Yu.A. Alexandrov et al., Phys. Part. Nucl. 30, 29 (1999).
47. H. Leeb, C. Teichtmeister, Phys. Rev. C 48, 1719 (1993).
48. Yu.A. Alexandrov, Phys. Rev. C 49, 2297 (1994).
References 49. S. Galster et al., Nucl. Phys. B 32, 221 (1971).
50. S. Platchkov et al., Nucl. Phys. A 510, 740 (1990).
1. E.M. Lyman, A.O. Hanson, M.B. Scott, Phys. Rev. 84, 51. M. Meyerho et al., Phys. Lett. B 327, 201 (1994).
626 (1951). 52. J. Becker et al., Eur. Phys. J. A 6, 329 (1999).
2. R. Hofstadter, R.W. McAllister, Phys. Rev. 98, 217 (1955). 53. M. Ostrick et al., Phys. Rev. Lett. 83, 276 (1999).
3. G.G. Simon et al., Nucl. Phys. A 333, 381 (1980). 54. C. Herberg et al., Eur. Phys. J. A 5, 131 (1999).
4. L. Andivahis et al., Phys. Rev. D 50, 5491 (1994). 55. D. Rohe et al., Phys. Rev. Lett. 83, 4257 (1999).
90 The European Physical Journal A
56. D. Rohe, these proceedings. 70. G. Hohler et al., Nucl. Phys. B 114, 505 (1976).
57. D.I. Glazier et al., Eur. Phys. J. A 24, 101 (2005). 71. W.R. Frazer, J.R. Fulco, Phys. Rev. Lett. 2, 365 (1959).
58. J. Bermuth et al., Phys. Lett. B 564, 199 (2003). 72. P. Mergell, U.G. Meissner, D. Drechsel, Nucl. Phys. A 596,
59. G. Warren et al., Phys. Rev. Lett. 92, 042301 (2004). 367 (1996).
60. R. Madey et al., Phys. Rev. Lett. 91, 122002 (2003). 73. H.W. Hammer, U.G. Meissner, D. Drechsel, Phys. Lett. B
61. R. Alarcon et al., contribution to the 16th International 385, 343 (1996).
Spin Physics Symposium (SPIN 2004), Trieste, Italy, 10- 74. H.W. Hammer, D. Drechsel, U.G. Meissner, Phys. Lett. B
16 Oct 2004. 586, 291 (2004).
62. R. Schiavilla, I. Sick, Phys. Rev. C 64, 041002 (2001). 75. H.W. Hammer, these proceedings
63. J. Friedrich, Th. Walcher, Eur. Phys. J. A 17, 607 (2003). 76. M.O. Distler (contact person) et al., Experiment MAMI
64. I. Passchier et al., Phys. Rev. Lett. 82, 4988 (1999). A1-2/2005.
65. H. Zhu et al., Phys. Rev. Lett. 87, 081801 (2001). 77. M.O. Distler, W. Heil, D. Rohe (contact persons) et al.,
66. T. Eden et al., Phys. Rev. C 50, 1749 (1994). Experiment MAMI A1-1/2005.
67. J. Friedrich, N. Voegler, Nucl. Phys. A 373, 219 (1982). 78. A. Liesenfeld et al., Phys. Lett. B 468, 20 (1999).
68. B. Krusche, Eur. Phys. J. A 26, 7 (2005). 79. R. Beck, these proceedings.
69. S. Kowalski, these proceedings. 80. H. Schmieden, these proceedings.
Eur. Phys. J. A 28, s01, 91 100 (2006)
DOI: 10.1140/epja/i2006-09-010-9 EPJ A direct
electronic only
/
Published online: 15 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Over the last decade accurate experiments at MAMI played an essential role to improve our
understanding of the nucleon to (1232) transition. Originally to a large extent motivated through intra
quark hyperne interactions anticipated in QCD-inspired quark models they showed that pionic degrees
of freedom are essential. The meson cloud is mainly responsible for the observed quadrupole excitation
strength and aects the magnetic dipole transition strength as well.
PACS. 13.60.Le Meson production 13.40.-f Electromagnetic processes and properties 14.20.Gk Baryon
resonances with S = 0
1 Introduction
At very high energy and momentum transfers in deep in-
elastic lepton scattering, proton and neutron reveal their
substructure of pointlike, almost massless spin-1/2 con-
stituents, the quarks, and of gluons as the exchange bosons
mediating the color force between them. In this regime
of asymptotic freedom Quantum Chromodynamics is well
established as the basic underlying theory of strong inter-
action. However, at momentum transfers corresponding to
the nucleons size, the nonlinear strong couplings prohibit
the solution of the QCD eld equations using perturbation
theory. Hence, basic properties such as mass, size and exci-
Fig. 1. Photoexcitation of the N transition. Top: M1
tation spectrum are only qualitatively [1] understood and
photon generating a spin-ip transition; bottom: E2 or C2 pho-
remain still a domain of models [2]. To prove that QCD ton introducing angular momentum L = 2.
provides also the correct theory at the connment scale is
the challenge of Lattice calculations, currently stepping
beyond quenched approximations [3] by realistic light
quark vacuum polarisation and chiral quark actions [4,5]. -nucleon channel, perfectly tting the present nominal
Analogously to atomic spectroscopy at the threshold to MAMI energy and thus most suited for a detailed study.
the era of quantum mechanics, baryon spectroscopy serves From the viewpoint of the inner quark dynamics the
today as a tool to improve our understanding of the inner nucleon to (1232) transition is very interesting. In the
dynamics of the nucleon. Generally, the high level density SU (6) symmetric constituent quark model it corresponds
of excited states in combination with their short lifetimes, to a pure spin-ip of one of the quarks, yielding the spin
and thus large natural widths, provides an annoying ex- 3/2 of the (1232). Electromagnetic excitation requires
perimental obstacle. However, there are two states which thus an M1 magnetic dipole photon as schematically de-
can be experimentally almost exclusively prepared, out- picted in g. 1 (top). Parity and angular momentum con-
standing in cross section and well separated in mass from servation would alternatively allow the absorption of an
their neighbours. The N (1535)S11 negative-parity partner L = 2 photon, coupling together with the nucleons spin
of the nucleon which selectively couples to the -nucleon 1/2 to J = 3/2, cf. g. 1 (bottom). However, this re-
nal state. Its mass puts it at the very edge of the energy quires L = 2 quadrupole components in one or both of
range accessible with MAMI B. Contrary, the decuplet the nucleons and deltas quark wave functions, in anal-
ground state (1232)P33 couples almost entirely into the ogy to the deuteron in nuclear physics. As in the latter
case, the quadrupole components can be associated with
a
e-mail: schmieden@physik.uni-bonn.de a spherical deformation of the system. They originate from
92 The European Physical Journal A
For photon energies up to 2000 MeV, g. 2 shows Fig. 3. Kinematics of pion electroproduction. The vectors of
the total absorption cross section of real photons on the the reaction plane are in the rest frame of the .
proton. The (1232) resonance exhibits almost isolated
around E = 340 MeV, whereas the second and third
resonance regions are composed of numerous overlapping In its rest frame, the recoiling hadronic system decays
resonances which, in the total absorption cross section, back to back into 0 and proton. and p denote the
cannot be separated. Due to their short life time, the di- pion and proton angles with respect to q in the hadronic
rect detection of the resonances is precluded. Since the cm frame. The reaction plane, which is given by l = pcm p ,
(1232) almost exclusively decays into pion and nucleon, n = q p , and t = n l, is tilted against the electron
cm
pion photo- and electro-production o the nucleon pro- scattering plane by the angle .
vide the major experimental tools for the investigation of In one photon exchange approximation the vefold dif-
its properties. ferential cross section
d5 d2 v
cm
= (1)
2 Kinematics and cross section of pion photo- dEe de d dcm
and electro-production factorizes into the virtual photon ux,
The kinematics of pion electroproduction are shown in E k 1
= , (2)
g. 3 at the example of the e+p e+p+ 0 reaction. The 2 2 E Q2 1
electron is scattered by the laboratory angle e and the
exchanged virtual photon transfers the dierence between and d2 v /dcm , the virtual photon cm cross section.
incoming and scattered electron energies and momenta, denotes the ne structure constant, k = (W 2 m2p )/2mp
= E E and q = k k , respectively. Q2 = q q > 0 the real photon equivalent laboratory energy for the exci-
represents the invariant mass of the virtual photon. The tation of the target with mass mp to the cm energy W ,
electron scattering plane is spanned by the unit vectors and = [1+(2|q|2 /Q2 ) tan2 2e ]1 the photon polarisation
z = q = q/|q|, y = k k , and x = y z. parameter.
Without target or recoil polarisation, the virtual pho-
ton cross section is given by [12]
0.6 d2 v
= [RT + L RL + c+ RLT cos +
dcm
0.5 RT T cos 2 + Pe c RLT sin ]. (3)
-N system -N system
parity
L -multipole J l -multipole
0 C0 1/2 1 S1 +
Fig. 4. Lowest-order graphs of single-pion electroproduction.
1/2 0 E0+ /S0+
E1/C1
3/2 2 E2 /S2
from non-resonant production (cf. g. 4). Therefore, to
1 1/2 1 M1
extract the small quadrupole admixture, a full multipole
M1 +
analysis would be desirable in analogy to -N scattering. 3/2 1 M1+
However, in pion production experiments with real and,
3/2 1 E1+ /S1+
in particular, virtual photons this is much more di cult E2/C2 +
to achieve, because more invariant amplitudes need to be 5/2 3 E3 /S3
independently determined. 2 3/2 2 M2
M2
5/2 2 M2+
2.1 Multipole decomposition
The response functions R of eq. (3) can be decomposed as electric/coulombic (E/C) or magnetic (M ). The pion
into pion multipoles of the nal state. In S and P wave multipoles AIl are characterized through their magnetic,
approximation this yields [12,13]: electric or scalar (longitudinal) nature, A = M, E, S (L),
the isospin, I, and the pion-nucleon relative angular mo-
RL = 2 [ |S0+ |2 + 4|S1+ |2 + |S1 |2 mentum, l . The coupling of l with the nucleon spin to
4e{S1+ S1 } + 2 cos e{S0+ (4S1+ + S1 )} the total angular momentum, J, is indicated by . Omis-
sion of the isospin index as in eqs. (4-8) indicates ampli-
+12 cos2 (|S1+ |2 + e{S1+ S1 }) ], (4)
tudes in the p 0 charge channel throughout this paper.
1
RT = |E0+ |2 + |2M1+ + M1 |2 In general, the unambiguous tagging of a resonance re-
2 quires the multipoles to be determined in the appropriate
1 isospin channel. However, at the resonance position of the
+ |3E1+ M1+ + M1 |2
2 (1232) the p 0 amplitudes provide a very good approx-
+2 cos e{E0+ (3E1+ + M1+ M1 )} imation to the separated isospin components. Both meth-
ods, simultaneous measurement of 0 and + production
+ cos2 |3E1+ + M1+ M1 |2 and sole 0 production, have been exploited.
1
|2M1+ + M1 |2
2 3 Experimental methods and results
1
|3E1+ M1+ + M1 |2 } , (5)
2 The sensitivity to the N (1232) quadrupole ampli-
3.1 Real photon experiments which provide a very high sensitivity to the EMR. In or-
der to exploit those experimentally, the cos 2 azimuthal
With real photon beams, i.e. Q2 = 0, the longitudinal modulation of the cross section needs to be determined.
parts of the cross section vanish. Of eqs. (4 8) thus only A detector with cylinder symmetry is ideally suited for
RT and RT T contribute and the cross section is often writ- this purpose. The rst experiments of the A2 collabora-
ten in the form [12] tion at MAMI consequently used the DAPHNE setup to
2 detect the recoil protons from the + p 0 + p reaction,
d2 d schematically depicted in g. 5 [14,15].
= [1 P cos 2], (11)
dcm dcm 0
Results of such measurements are shown in g. 6. Ex-
d2 ploiting the cleanly measured azimuthal modulation (left
where [ d cm ]0denotes the unpolarised cross section and
part of gure) the photon beam asymmetries (right part)
P is the degree of linear polarisation of the photon can be determined as a function of polar angle over the
beam. The photon-beam asymmetry represents the ra- entire energy region of the (1232) resonance. In conjunc-
tio RT T /RT and thus enters the cross section with the tion with the simultaneously measured dierential cross
cos 2 modulation characteristic for the directional sensi- sections it is possible to perform a detailed mulipole anal-
tivity against the plane of linear polarisation. This plane ysis. Results are depicted in g. 7 [15,16]. Furthermore,
is intrinsically xed by kinematic constraints in electron including the results of the reaction + p + + n with
scattering, usually the detection of the nal-state electron linearly polarised photon beam enables an isospin decom-
under a certain angle. Contrary, real photon beams from position. At the resonance position it provides EMR re-
electron bremsstrahlung are a homogeneous superposition sults in agreement with the 0 channel alone. As nal re-
of all polarisation directions. The net polarisation conse- sult [15] the electric quadrupole to magnetic dipole ratio
quently vanishes.
To obtain linear polarised photon beams it is necessary
to x the electron scattering plane in the bremsstrahlung
process. This can be achieved by o-axis tagging or, as
exploited at MAMI, through coherent bremsstrahlung o
a diamond crystal. Under certain kinematic conditions de-
pending on the crystal alignment relative to the electron
beam the crystal lattice, similar to the Mossbauer eect,
coherently takes the bremsstrahlung recoil instead of indi-
vidual nuclei. This generates intensity peaks in the other-
wise 1/E distribution of the bremsstrahlung spectrum.
Within the energy region of the coherent peak the crys-
tal orientation determines a particular electron scattering
plane. Hence, the photon beam is linearly polarised and
the degree of linear polarisation is related to the intensity
excess over the incoherent spectrum. Fig. 6. Left: Measured relative cross section of the reaction
According to eq. (11), linear polarised beams with + p 0 + p using linearly polarised photon beam as a
P = 0 enable the contribution of the beam asymmetry function of the azimuthal angle at xed polar angle cm =
to the observed cross section. Its size is determined by 90 . Right: Polar angle dependence of the extracted photon
interferences of the type |M1+ |2 + 6 e{E1+ M1+ } + beam asymmetry.
e{E1+ M1+ }
= (2.5 0.1stat 0.2syst ) % (12)
|M1+ |2
was obtained.
The beam asymmetry had also been extracted using
the large acceptance TAPS photon detector, covering the
full polar angular range for the 0 decay photons [17].
Despite the coplanar detector arrangement, could be
determined through the deliberate rotation of the photon
polarisation plane.
via
have been measured [21]. They are shown in g. 9 along
c+ RLT with dierent calculations. The prediction of Sato and
LT (cm
0 ) = . (14) Lee [22] describes the asymmetries quite well, to a lesser
RT + L RL + RT T
96 The European Physical Journal A
tion [37]:
Fig. 12. Beam-helicity asymmetry LT measured in 0 pro-
duction with longitudinally polarised electron beam. The dot-
0 Px = Pe 2L (1 )RLT t
, (22)
ted curve represents the original MAID2003 calculation [23].
The dashed-dotted and dashed curves are the results of the 0 Py = 2L (1 + )RLT ,
n
(23)
dynamical models of Sato-Lee [22] and Kamalov-Yang [25], re-
spectively. The full curve is the MAID re-t of ref. [21]. Errors 0 Pz = Pe 1 2 RTl T . (24)
are purely statistical.
According to g. 13 the axes are dened relative to the
electron scattering plane and the virtual photon direction.
This is the natural choice, since in parallel kinematics the
with this measurement. However, rescaled by a factor 0.75 recoil polarisation is completely determined by the angular
MAID described the asymmetry very well. The later re- momentum transfer from the photon eld. Decomposition
t of real and imaginary parts of the S1+ and S0+ 0 - up to S and P partial waves,
amplitudes in MAID2003 on basis of the MAMI LT and
LT data provided a very satisfactory description of the 0 Px = Pe c e{(4S1+ + S1 S0+ )
polarisation data, cf. the full curve in g. 12, and of the (M1+ M1 E0+ + 3E1+ )}, (25)
unpolarised measurement discussed in the preceding sec-
0 Py = c+ m{(4S1+ + S1 S0+ )
tion as well [21]. This underlines the absolute need to un-
derstand the physical background amplitudes before the (M1+ M1 E0+ + 3E1+ )}, (26)
small quadrupole contributions can be reliably extracted. 0 Pz = Pe c0 |M1+ |2 + |M1 |2 + 9|E1+ |2 + |E0+ |2
Ideally, a complete experiment with respect to a multipole + e{6E1+
(M1+ M1 ) 2M1+
M1
decompostion is required.
2E0+ (M1+ M1 + 3E1+ )} , (27)
ponent. A separation of Px , Py and Pz is achieved since Fig. 14. CMR from recoil polarisation (open diamond) [38]
x and z components are odd under beam-helicity rever- in comparison to unpolarised measurements [18, 19, 26, 28,
29, 44, 45] at low Q2 . The curves show model calcula-
sal, while Py is even (cf. eqs. (22) (24)). At the energy
tions MAID2003 [23] (solid), DMT2001 [25] (dashed) and
of the (1232) resonance and a momentum transfer of
Sato/Lee [22] (dashed dotted).
Q2 = 0.121 (GeV/c)2 the three polarisations were simul-
taneously measured to
1
momentum conservation [46]:
Px /Pe = (11.4 1.3stat 1.4syst ) %, (29) 1
2x + 2y = z (1 z ). (33)
Py = (43.1 1.3stat 2.2syst ) %, (30) L
Pz /Pe = (56.2 1.5stat 2.6syst ) %. (31) The reduced polarisations are dened by
t
1 RLT
x = Px = , (34)
0
Pe c R T + L R L
Based on the Px result, the CMR n
1 RLT
y = Py = , (35)
c+ R T + L R L
{S1+ M1+ } 1 RTl T
= (6.4 0.7stat 0.8syst ) % (32) z = Pz = . (36)
|M1+ |2 Pe c0 R T + L R L
The consistency relation seems hardly fullled by the
was extracted using the MAID parametrisation [38]. measured recoil polarisations. The experimental result for
Within the errors this agrees with the alternative method eq. (33) is
of extraction using Px /Pz , which experimentally provides
3.9 0.4stat 0.4syst = 7.9 0.7stat 1.2syst . (37)
the advantage that the magnitudes of both analysing
power and beam polarisation drop out. Contrary, the ra- Despite the non-linear error propagation on the r.h.s. of
tio is aected by a larger uncertainty in spin precession of eq. (33), the probability for such a nding is only a few
the z-component compared to Px . To rst order the latter percent. While MAID2000 fullls the consistency rela-
would not precess at all in a homogeneous vertical-bend tion, there is a discrepancy with the Py measurement.
dipole eld. The MAID-ret mentioned above slightly improves this
0 situation as shown in g. 15, where the MAMI recoil po-
The result for the CMR is depicted in g. 14 along
larisation data [38] are compared to the MAID versions
with unpolarised measurements. Keeping in mind that
0 2000, 2003 and the re-t. Nevertheless, the role of non-
the large negative CMR of ref. [26] (full triangle tip resonant background, which particularly shows up in Py ,
down) is practically excluded by the S0+ results discussed still remains unresolved. Another measurement of Py [48]
in sect. 3.2.1, very convincing agreement is observed be- is in agreement with the MAMI data but has a large er-
tween the various measurements. A recent very high statis- ror. Unfortunately, ref. [43] gives no explicit values of the
tics measurement of the angular distributions of recoil measured polarisations.
polarisation [43] found, just outside the range of g. 14,
0 1
CMR = (6.610.18) % at Q2 = 1 GeV/c2 . This agree- Despite the occurence of L in eq. (33), the polarisation
ment extends also to the positive S0+ /M1+ ratio deter- relation is nevertheless frame independent. The longitudinal
mined from LT . polarisation parameter is also contained in the denition of
the factors c of the reduced polarisations and thus could be
The three polarisation components measured in paral- eliminated. A similar relation holds in elastic electron nucleon
lel kinematics are model independently related by angular scattering, where y vanishes [47].
H. Schmieden: Photo- and electro-excitation of the -resonance at MAMI 99
5
-20 70
0 -25
-30 65
(%)
(%)
(%)
-5 -35
Pzsp / Pe
Pxsp / Pe
60
Pysp
-40
-10
-45
55
-50
-15
-55 50
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4
Q2 (GeV2/c2) Q2 (GeV2/c2) Q2 (GeV2/c2)
Fig. 15. Components of recoil polarisation measured at the energy of the (1232)-resonance in the reaction p(e, e p) 0 [38].
Curves show the MAID versions 2000 (full), 2003 (dashed), and the re-t (dashed-dotted) on basis of LT and LT [21] discussed
in the text. Note the suppressed zero middle and right.
In principle, from the reduced polarisations it is possi- Obviously, pionic degrees of freedom which in a mi-
ble to determine the ratio of longitudinal to transverse re- croscopic picture would be responsible for e.g. the mS0+
sponse, RL /RT , without the need of a classical Rosenbluth contribution (g. 10) play an important role. This is
separation. According to ref. [46] this can be achieved in made particularly transparent within the dynamical mod-
three dierent ways, using either els [24,22]. Such calculations can be split into the so-called
bare and dressed parts, which represent the nucleon/delta
only the longitudinal component, z (cf. also [49]), core, and the pion cloud, respectively. This is also at-
the quadratic sum of the transverse components, tempted within dispersion relation approaches [50]. While,
2x + 2y , or in general, such a separation suers from an unitary am-
all three reduced polarisations. biguity [51], the models yield consistent results. E.g., the
However, as a re ection of the almost violated consistency full model of ref. [24] describes the data while the bare cal-
relation the results obtained from the MAMI measure- culation yields a very small and positive CMR. Moreover,
3/2
ments vary signicantly. The smallest value, RL /RT = the full calculation yields an M1+ amplitude in agreement
(4.7 0.4stat 0.6syst ) %, is extracted from the quadratic with the experimental data. In contrast, the bare calcu-
3/2
sum 2x + 2y of the transverse reduced polarisations and lation gives only about two thirds of M1+ and a CMR
the largest one, RL /RT = (12.2 +1.7 +2.9
1.6stat 2.7syst ) %, from
which is an order of magnitude too small. This is simi-
z alone. This presently prohibits a reliable extraction lar to quark model calculations without pion degrees of
of RL /RT but stresses the importance of a simultaneous freedom [52,53], which also underestimate the quadrupole
measurement of all polarisation components with further strength by a factor of ten and get only about 60 % of the
improved accuracy. experimental M1+ .
Despite the yet unsatisfactory statistics of dynamic lat-
tice calculations [54], those are in qualitative agreement
with the experimental results and hence further support
4 Interpretation the important role of the pion cloud.
As outlined in the introduction the quadrupole transi-
The MAMI experiments towards the quadrupole strength tion strength can be interpreted in terms of a deformation
in the N (1232) transition yield very consistent re- of the baryons involved. Buchmann and Henley nd oppo-
sults. The rst order physical background contributions site deformations of equal strength for nucleon and [55].
like mS0+ are now much better under control. Remain- Also the quadrupole moments for the nucleon core and
ing uncertainties and inconsistencies seem related to the the pion cloud are of opposite sign. However, the core ap-
contribution of higher partial waves especially in the imag- pears almost spherical and the deformation due to the pion
inary parts of interferences as in LT and Py . Though cloud is an order of magnitude stronger, very similar to
important for our understanding of the non-resonant pro- the observation within the dynamical model. Buchmann
cesses, they are very much suppressed in the discussed ob- and Henley quote total quadrupole moments in the range
servables with high sensitivity to EMR and CMR. At least of Qnucl = (0.113 . . . 0.5) fm2 for nucleon (upper sign)
in the vicinity of the photon point, a reliable extraction of and (lower sign). It is interesting to note that, relative
the EMR and CMR is possible with a remaining relative to the size of the objects, the corresponding deformation
model uncertainty of the order 10 %. Thus it is evident
that the experimental results are an order of magnitude Qnucl
larger than expected from the quark model calculations. 2 > 0.043
rN
of the nucleon
100 The European Physical Journal A
/
Published online: 26 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Parity-violating electron scattering has been a very useful tool for probing the structure of
neutral currents and providing detailed information on electroweak form factors. A pioneering SLAC mea-
surement in the mid-70s provided an important early test of the Standard Model. Modern electron acceler-
ators provide high-intensity (> 100 A), CW beams with polarizations as high as 85%. Experiments such
as SAMPLE, A4, HAPPEX and G0 have exploited these capabilities and obtained new information on
electroweak strange form factors in the Q2 range of 0.1 1.0 (GeV/c)2 . That activity continues. Other ex-
periments are designed to provide stringent tests of the Standard Model. E-158 at SLAC recently measured
the weak charge of the electron. Qweak is a challenging new experiment at JLAB which is designed to mea-
sure the weak charge of the proton. This will probe for physics beyond the Standard Model corresponding
to energy scales of more than 5 TeV.
PACS. 12.15.-y Electroweak interactions 14.20.Dh Protons and neutrons 21.10.Gv Mass and neutron
distributions
GF
AP V =
4 2
p Z
GE GE + GpM GZ p
M (1 4sin W )GM GA
2 e
. (1)
(GpE )2 + (GpM )2
= Q2 /4Mp2 , (2)
1
= 1 + 2(1 + )tan2 (e /2) (3)
and
= (1 2 ) (1 + ) . (4)
1. SAMPLE
SAMPLE [4] at MIT-Bates was the rst experiment to use
parity violation as a probe for strange quarks in the pro-
ton. A longitudinally polarized electron beam of 200 MeV
was incident on a liquid hydrogen or deuterium target.
The polarized electrons were produced using a bulk GaAs
crystal resulting in an average polarization of 36%. The
linac produces a pulsed beam of 25 s duration at a rep-
etition rate of 600 Hz. The beam current was 40 A. The
helicity of the beam is changed randomly pulse-by-pulse
and in addition a half-wave plate can be inserted to change Fig. 2. Uncertainty bands of GsM vs. GeA at Q2 = 0.1 (GeV/c)2
the overall sign of the helicity. for the SAMPLE experiment in both hydrogen and deuterium.
The scattered electrons were detected in a large solid Also shown is the uncertainty band of the theoretical expecta-
angle (1.5 sr) air Cerenkov detector spanning angles be- tion for GeA .
tween 130 and 170 . At backward angles SAMPLE is
mostly sensitive to GsM and GeA . Figure 1 shows one of
the 10 detector modules which are placed A symmetri- was critical to the success of the experiment. Several feed-
cally about the beam axis. The Cherenkov light is focused back systems were used to minimize such eects. These
by an ellipsoidal mirror unto a phototube. The integrated included energy, beam position, angle, and intensity. All
light is proportional to the scattered electron rate of about parity experiments implement similar feedback controls.
108 s1 in a beam pulse. Experiments were carried out at 200 MeV on both hy-
A shutter located in front of the phototube could be drogen and deuterium targets. In addition a measurement
closed providing a measurement of the background origi- at 125 MeV was also made on a deuterium target. The two
nating from neutrons and charged particles. Tight control targets in principle allow a separation of GsM and GeA . The
of helicity-correlated eects on the properties of the beam results for both hydrogen and deuterium are summarized
S. Kowalski: Parity violation in electron scattering 103
where
FW (Q2 ) = (1 4 sin2 W )Fp (Q2 ) N Fn (Q2 ), (10)
and
F (Q2 ) = ZFp (Q2 ) (11)
A measurement of AP V to 3% would provide a mea-
Fig. 5. Layout of the G0 spectrometer system. Shown are the
superconducting toroidal magnet and the segmented scintilla-
surement of Rn to 1%. An experiment has been approved
tor detector array. at JLAB to carry out such a measurement. An 850 MeV,
50 A polarized beam would be incident on a lead target
sandwiched between diamond sheets for cooling. Electrons
measures a dierent Q2 range. The beam duty factor was scattered through 60 would be detected in the pair of Hall-
reduced to 6% allowing timing measurements to be used to A high-resolution spectrometers. At Q2 = 0.01 (GeV/c)2
separate background from the recoiling protons. The de- the parity-violating asymmetry, AP V = 0.5 ppm.
tector operates in counting mode. The A4 experiment at This is a very challenging experiment. It will require
Mainz also operates in counting mode. All other electron control of helicity correlated systematics to much better
parity violation experiments operate in integrating mode. than 15 ppb. It appears to be feasible.
The results [8] are shown graphically in g. 4 together
with the results from other experiments. The results in-
dicate a non-zero, Q2 -dependent, strange quark contribu- 2.3 Standard Model tests
tion. They cover a much broader Q2 range than the other
experiments. The uncertainties for the higher Q2 points Parity violating electron scattering has been used to test
are quite large as a result of signicant unanticipated back- the Standard Model (SM) since the pioneering SLAC ex-
ground contributions. periment in the mid-70s. Collider experiments at the Z-
pole provide our best measure of the weak mixing angle,
G0 is planning to make backward angle measurements
in 2006. At backward angles, scattered electrons are de- sin2 W . There remains much interest in exploring the run-
tected instead. Each measurement will be at a single Q2 . ning of sin2 W from the Z-pole to Q2 = 0.
They expect to make the rst measurements at approxi- At Q2 = 0, atomic parity experiments give results
mately 0.2 and 0.6 (GeV/c)2 . consistent with the running of sin2 W . Other experi-
ments testing the SM include 12 C(e, e), 9 Be(e, e), NuTeV,
Moeller scattering and Qweak . All parity violating experi-
ments can be described in terms four fundamental quark
2.2 Neutron densities
coupling constants,
12
1. C / 9 Be
12
C is a spinless and isoscalar nucleus and elastic electron
scattering is described by a single form factor. The parity-
violating asymmetry may be written at the tree level [10,
11] as 1
3 2
AP V = 2 GF Q ( 2 ) . (18)
A parity violation experiment was carried out at MIT-
Bates [12] on 12 C. The results are shown in g. 6 to-
gether with the results of the SLAC experiment [13] on
deuterium. The results of both experiments are consistent
with the predictions of the SM.
2. E-158
A recent experiment testing the SM was E-158 at SLAC.
It is a purely leptonic process involving Moeller scatter- Fig. 8. Calculated running of the weak mixing angle in the
ing. The goal of the experiment was to measure the weak Standard Model. Data points are from the atomic parity vio-
charge of the electron, e = g sin W , using parity violation. lation experiment on Cs, the NuTeV experiment, the Moeller
In this experiment 50 GeV electrons were incident on experiment (E-158) at SLAC and from experiments at the Z0 -
a liquid hydrogen target. At Q2 = 0.027 (GeV/c)2 the pole. Also shown are the anticipated error bars for Qweak .
parity violating asymmetry is expected to be 150 ppb. A
quadrupole spectrometer was used to focus the scattered
Moeller electrons while at the same time defocusing e-p tons weak charge, QpW = 1 4 sin2 W , to the highest
scattering events. The ux of scattered electrons was in- precision possible.
tegrated for each beam burst. The experiment yielded an The SM makes a rm prediction of QpW based on the
asymmetry, running of the weak mixing angle sin2 W from the Z0 -
pole to low energies, corresponding to a 10 eect in our
AP V = 175 30 20 ppb. (19)
experiment. Fig. 8 shows the SM prediction for sin2 W
The extracted weak mixing angle is totally consistent together with existing data and the expected precision for
with predictions. The result [14] is shown plotted in g. 8. this experiment. This parity violating experiment is in the
semi-leptonic sector. This is in contrast to E-158 which is
3. Qweak in the purely leptonic sector.
A major new initiative is under development at JLAB. The measurement will be carried out using a 1.2 GeV
The goal of the Qweak experiment is to measure the pro- electron beam at a scattering angle of 9 and a momentum
106 The European Physical Journal A
transfer Q2 = 0.03 (GeV/c)2 . The 180 A polarized beam Such demanding measurements have been made possi-
will be incident on a 35 cm liquid hydrogen target. An ble by important advances in accelerator technology. We
eight sector toroidal spectrometer is being constructed for now have high intensity CW beams with beam polariza-
this measurement. The scattered electrons will be detected tion of 85%. Control of helicity correlated beam proper-
by quartz Cherenkov detectors operating in integrating ties allows measurement of asymmetries to an accuracy
mode. A schematic layout of the spectrometer is shown in approaching 10s ppb.
g. 7.
The measurement will take 2200 hours and will deter-
mine the protons weak charge with 4% statistical accu- References
racy. Figure 8 shows the projected quality of the Qweak
results in the context of other existing data. 1. R.D. McKeown, Phys. Lett. B 219, 140 (1989).
2. D. Keyslan, A. Manohar, Nucl. Phys. B 310, 527 (1988).
3. M.J. Musolf et al., Phys. Rep. 239, 1 (1994).
3 Summary 4. D.T. Spangle et al., Phys. Lett. B 583, 79 (2004).
5. K.A. Aniol et al., Phys. Rev. C 69, 065501 (2004).
Parity-violating high-energy electron scattering is an im- 6. The HAPPEX Collaboration (K.A. Aniol et al.), Phys.
portant probe of nucleon and nuclear structure. Several Lett. B 635, 275 (2006), nucl-ex/0506011.
physics areas are currently under investigation. There 7. The HAPPEX Collaboration (K.A. Aniol et al.), Phys.
Rev. Lett. 96, 022003 (2006), nucl-ex/0506010.
have been many experiments investigating the importance
8. D.S. Armstrong et al., Phys. Rev. Lett. 95, 092001 (2005).
of strange quarks to the structure of the proton. Results
9. T.W. Donnelly et al., Nucl. Phys. A 503, 589 (1989).
to date indicate that the contribution of strange quarks 10. G. Feinberg, Phys. Rev. D 12, 3575 (1975).
to nucleon structure are relatively small or vanishing. A 11. J.D. Walker, Nucl. Phys. A 285, 345 (1977).
proposed experiment on Pb would measure the neutron 12. P.A. Souder et al., Phys. Rev. Lett. 65, 694 (1990).
radius to 1%. Qweak is a challenging Standard Model test 13. C.Y. Prescott et al., Phys. Lett. B 84, 524 (1979).
which probes 5 TeV energy scales. 14. P.L. Anthony et al., Phys. Rev. Lett. 95, 081601 (2005).
Eur. Phys. J. A 28, s01, 107115 (2006)
DOI: 10.1140/epja/i2006-09-012-7 EPJ A direct
electronic only
/
Published online: 29 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. A measurement of the weak form factor of the proton allows a separation of the strangeness
contribution to the electromagnetic form factors. The weak form factor is accessed experimentally by the
measurement of a parity violating (PV) asymmetry in the scattering of polarized electrons on unpolarized
protons. We performed such measurements with the setup of the A4-experiment at the MAMI accelerator
facility in Mainz. The role of strangeness in low energy nonperturbative QCD is discussed. The A4-
experiment is presented as well as the results on the strangeness form factors which have been measured
at two Q2 -values. The plans for backward angle measurements at the MAMI facility are presented.
PACS. 12.15.-y Electroweak interactions 11.30.Er Charge conjugation, parity, time reversal, and other
discrete symmetries 13.40.Gp Electromagnetic form factors 25.30.Bf Elastic electron scattering
Moellerpolarimeter (A1)
P=80%, I=20A
of three terms, ALR (ep) = AV + As + AA . high power transmission
MAMI liquid Compton polarimeter
E=855MeV
GpE GnE + GpM GnM E/E=10-6
Compton- hydrogen
E,
and beam dump
AV = aeq (1 4eq sZ )
2
, (1) laser target
(GpE )2 + (GpM )2 back scatter
polarimeter
e-
luminosity
Gp Gs + GpM GsM e-
As = aeq E p E2 , (2)
e-
(GE ) + (GpM )2 e- E,
(1 4s2Z ) 1 2 (1 + )GpM GpA calorimeter:
AA = a . (3)
(GpE )2 + (GpM )2 1022 PbF2-crystals
AV arises from the Z0 coupling to the proton vector cur- Fig. 2. Measurement principle of the A4 experiment. The po-
rent and contains the electromagnetic nucleon form fac- larized electrons from the source are accelerated in the MAMI
accelerator to a maximum energy of 855 MeV. Scattered elec-
tors Gp,n
E,M . A possible strangeness contribution to the trons from the 10 cm hydrogen target are detected in the homo-
proton electromagnetic vector form factors has been sep- geneous 1022 channel PbF2 -Cherenkov calorimeter. The sen-
arated into As . The coupling to the proton axial cur- sitive measurement and stabilization of all electron beam pa-
rent is presented by AA and contains the neutral cur- rameters is crucial for the sensitivity of the experiment.
rent weak axial form factor GpA . e is the scattering an-
gle of the electron in the laboratory and Q2 the negative
square of the four momentum transfer. = Q2 /(4Mp2 ) and violating cross section asymmetry in the scattering of po-
= [1+2(1+ ) tan2 (e /2)]1 represent purely kinematical larized electrons o unpolarized protons. It is complemen-
factors with Mp the proton mass. G and represent the tary to other experiments since for the rst time counting
Fermi coupling constant as derived from muon decay and techniques have been used in a PV electron scattering ex-
the ne structure constant respectively. a denotes the fac- periment. Possible systematic contributions to the exper-
tor (G Q2 )/(4 2). In order to average A0 = AV + AA imental asymmetries and the associated uncertainties are
over the acceptance of the detector and the target length, of a dierent nature as compared to previous experiments,
we take values for the electromagnetic form factors Gp,nE,M which use analogue integrating techniques. Figure 2 shows
from a parametrization (version 1, page 5) by Friedrich the measurement principle of the A4 experiment.
and Walcher [19] and assign an experimental error of 3 % The polarized 570.4 and 854.3 MeV electrons were pro-
to GpM and GpE , 5 % to GnM , and 10 % to GnE . For evalu- duced using a strained layer GaAs crystal that is illu-
ating A0 , electromagnetic internal and external radiative minated with circularly polarized laser light [24]. Aver-
corrections to the asymmetry and energy loss due to ion- age beam polarization was about 80 %. The helicity of
ization in the target have been calculated and they reduce the electron beam was selected every 20.08 ms by set-
the expected asymmetry for our kinematics by 1.3 %. ting the high voltage of a fast Pockels cell according to
Electro-weak quantum corrections have been applied a randomly selected pattern of four helicity states, either
in the M S renormalization scheme according to [20] (+P P P + P ) or (P + P + P P ). A 20 ms time
and are contained in the factors eq , with s2Z = window enabled the histogramming in all detector chan-
sin2 W (MZ )M S = 0.23120(15) [21]. The electro-weak nels and an integration circuit in the beam monitoring and
quantum corrections to AA [22] are applied and absorbed luminosity monitoring systems. The exact window length
in the value of GpA . was locked to the power frequency of 50 Hz in the labora-
The largest contribution to the uncertainty of A0 tory by a phase locked loop. For normalization, the gate
comes from the uncertainty in the axial form factor GpA , length was measured for each helicity. Between each 20 ms
the electric form factor of the proton GpE , and the mag- measurement gate, there was an 80 s time window for the
netic form factor of the neutron GnM . For our experimental high voltage at the Pockels cell to be changed. The inten-
program at backward angles, the uncertainty stemming sity I = 20 A of the electron current was stabilized to
from GpA will be more important in the separation of GsE better than I/I 103 . An additional /2-plate in the
and GsM . Using a dierent target like deuterium will allow optical system was used to rotate small remaining linear
polarization components and to control the helicity corre-
to separate GpA , GsE and GsM . The parity admixture in the
lated asymmetry in the electron beam current to the level
ground state of deuterium is negligible [23].
of < 10 ppm in each ve minute run.
From the source to the target, the electron beam de-
4 The A4 experimental setup and analysis velops uctuations in beam parameters such as position,
energy and intensity which are partly correlated to the
The A4 experiment at MAMI has been developed and reversal of the helicity from +P to P . We have used a
build in order to measure a small (order ppm) parity- system of microwave resonators in order to monitor beam
110 The European Physical Journal A
x 102
1800
Counts
1600
1400
Elastic Peak
1200 Elastic Cut
1000
800 Elastic Cut
600
400
200
0
0 50 100 150 200 250
ADC-Channel 32
Fig. 6. The dashed histogram shows a raw energy spectrum
of accepted particles from the hydrogen target as read directly
from the hardware memory of the readout electronics of the
lead uoride calorimeter. For the solid black curve, the raw
spectrum has been corrected for the dierential nonlinearity
of the ADC, i.e. for measured variations of the ADC channel
width. The elastic scattering peak position, the 0 -production
threshold and the -resonance position are indicated as well
as the lower and upper cut position for the extraction of NeR
and NeL as described in the text.
Fig. 5. Design drawing of the A4 readout electronics. The
upper part contains the analog sum, trigger and veto circuits
together with the digitization. In the lower part, the histogram-
ming and the VMEbus access is done. The system is about due to dead time were investigated by varying the beam
3.5 m high. current. We calculate the raw normalized detector asym-
metry as Araw = (NeR /R NeL /L )/(NeR /R + NeL /L ).
The possible dilution of the measured asymmetry by back-
the 1022 channels were installed. The particle rate within ground originating from the production of 0 s that sub-
the acceptance of this solid angle was 50 106 s1 . sequently decays into two photons where one of the pho-
Due to the short dead time, the losses due to double hits tons carries almost the full energy of an elastic scattered
in the calorimeter were 1 % at 20 A. This low dead time electron was estimated using Monte Carlo simulations to
is only possible because of the special readout electronics be much less than 1 % and is neglected here. The largest
employed. The signals from each cluster of 9 crystals were background comes from quasi-elastic scattering at the thin
summed and integrated for 20 ns in an analogue summing aluminum entrance and exit windows of the target cell. We
and triggering circuit and digitized by a transient 8-bit have measured the aluminum quasi-elastic event rate and
ADC. There was one summation, triggering, and digitiza- calculated in a static approximation a correction factor for
tion circuit per crystal. The energy, helicity, and impact the aluminum of 1.030 0.003 giving a smaller value for
information were stored together in a three dimensional the corrected asymmetry.
histogram. Neighboring crystals have to go to neighboring
electronics channels in the electronics resulting a ring Corrections due to false asymmetries arising from he-
shape. Analogue summation and digitization has been licity correlated changes of beam parameters were applied
galvanically separated from histogramming and VMEbus on a run by run basis. The analysis was based on the ve
access. Figure 5 shows a design drawing of the fast A4 minute runs for which the counted elastic events in the
experiment electronics. PbF2 detector were combined with the correlated beam
parameter and luminosity measurements. In the analysis
Figure 6 shows an energy spectrum of scattered par-
we applied reasonable cuts in order to exclude runs where
ticles. The number of elastic scattered electrons is deter-
the accelerator or parts of the PbF2 detector system were
mined from this histogram for each detector channel by in-
malfunctioning. The analysis is based on a total of 7.3106
tegrating the number of events in an interval from 1.6 E
histograms corresponding to 4.8 1012 elastic scattering
above pion production threshold to 2.0 E above the elas-
events for the 854.3 MeV data and 4.8106 histograms cor-
tic peak in each helicity histogram, where E is the en-
responding to 21013 elastic events for the 570.4 MeV data.
ergy resolution for nine crystals. These cuts ensure a clean
separation between elastic scattering and pion production For the correction of helicity correlated beam param-
or -excitation which has an unknown PV cross section eter uctuations we used multi-dimensional linear regres-
asymmetry. The linearity of the PbF2 detector system sion analysis using the data. The regression parameters
with respect to particle counting rates and possible eects have been calculated in addition from the geometry of the
112 The European Physical Journal A
20
Asymmetry [ppm]
15
10
5
0
-5
-10
out out out out out out out out
-15 out
in in in in in in in in
-20
0 2 4 6 8 10 12 14 16
Sample No.
5
Asymmetry (ppm)
4
3
2
1
0
-1
-2
-3
-4 out in out in in out in out
-5
1 2 3 4 5 6 7 8
Sample No.
Fig. 7. The top plot shows the data samples of 854.3 MeV
data with the /2-plate in and out. The lower plots represents
the data sample for the 570.4 MeV data with the /2-plate in Fig. 8. Top: The solid line represents all possible combinations
and out as described in the text. of GsE + 0.225GsM as extracted from the work presented here at
a Q2 of 0.230 (GeV/c)2 . The densely hatched region represents
the 1- uncertainty. The recalculated result from the HAPPEX
precisely surveyed detector geometry. The two dierent published asymmetry at Q2 of 0.477 (GeV/c)2 is indicated by
methods agree very well within statistics. the dashed line, the less densely hatched area represents the as-
sociated error of the HAPPEX result. Bottom: The solid lines
The experimental asymmetry is normalized to the elec- represent the result on GsE + 0.106GsM as extracted from our
tron beam polarization Pe to extract the physics asym- new data at Q2 = 0.108 (GeV/c)2 presented here. The hatched
metry, Aphys = Aexp /Pe . We have taken half of our data region represents in all cases the one--uncertainty with sta-
with a second /2-plate inserted between the laser system tistical and systematic and theory error added in quadrature.
and the GaAs crystal. This reverses the polarization of The dashed lines represent the result on GsM from the SAM-
the electron beam and allows a stringent test of the un- PLE experiment [31]. The dotted lines represent the result of
derstanding of systematic eects. The eect of the plate a recent lattice gauge theory calculation for s [32]. The boxes
can be seen in g. 7: the observed asymmetry extracted represent dierent model calculations and the numbers denote
from the dierent data samples changes sign, which is a the references.
clear sign of parity violation if, as in our case, the target
is unpolarized. 5 Conclusion
Our measured result for the PV physics asymmetry in
the scattering cross section of polarized electrons on unpo- From the dierence between the measured ALR (ep) and
larized protons at an average Q2 value of 0.230 (GeV/c)2 the theoretical prediction in the framework of the Stan-
is ALR (ep) = (5.44 0.54stat 0.26syst ) ppm for the dard Model, A0 , we extract a linear combination of
854.3 MeV data [29] and ALR (ep) = (1.36 0.29stat the strange electric and magnetic form factors for the
0.13syst ) ppm for the 570.4 MeV data [30]. The rst error 570.4 MeV data at a Q2 of 0.108 (GeV/c)2 of GsE
represents the statistical accuracy, and the second rep- 0.106 GsM = 0.071 0.036. For the data at 854.3 MeV
resents the systematical uncertainties including beam po- corresponding to a Q2 value of 0.230 (GeV/c)2 we extract
larization. The absolute accuracy of the experiment repre- GsE + 0.225 GsM = 0.039 0.034. Statistical and system-
sents the most accurate measurement of a PV asymmetry atic error of the measured asymmetry and the error in the
in the elastic scattering of longitudinally polarized elec- theoretical prediction of A0 been added in quadrature. In
trons on unpolarized protons. g. 8 the results for the 570.4 MeV data are displayed.
F.E. Maas: Parity-violating electron scattering at the MAMI facility in Mainz (A4) 113
Beam
norm
-----------------------
norm pol1 1 raw
Kurtosis -2.86987e-01
conditions:
pol0 1 T=300
raw s, I=19.9 muA
conditions:
norm
1 raw
Kurtosis -6.04778e-01
pol0 1 T=300
raw s, I=19.9 muA
conditions:
norm T=300 s, I=19.9 muA
pol0 1 raw
conditions:
norm
1 raw
Kurtosis -5.11082e-01
pol0 1 T=300
raw s, I=19.9 muA
4000 4000 4000 4000 From Parity Violation to Hadronic Structure and More,
2000
0
2000
0
2000
0
2000
0
Proceedings of PAVI2004, Eur. Phys. J. A 24, s02 (2005).
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
7. T.R. Donoghue, E. Golowich, B.R. Holstein, Dynamics of
Run 28256: Chan 922 pol1 ProjX EntriesMean 1.80404e+01
Normal: 2.59947e+06
Run 28256: Chan 923 pol1 ProjX EntriesMean 1.93011e+01
Normal: 2.02822e+06
Run 28256: Chan 924 pol1 ProjX EntriesMean 1.78771e+01
Normal: 1.42570e+06
Beam
EntriesIntegral
norm
Coinc: 5.00314e+05
1.29944e+06
-----------------------
norm pol1 1 raw
Kurtosis -4.73486e-01
conditions:
pol0 1 T=300
raw s, I=19.9 muA Beam
EntriesIntegral
Coinc: 3.88322e+05
1.01361e+06
-----------------------
norm pol1 1 raw
Kurtosis -5.08935e-01
conditions:
norm pol0 1 T=300
raw s, I=19.9 muA Beam
EntriesIntegral
norm pol1
Coinc: 3.30970e+05
1 raw
7.12172e+05
-----------------------
Kurtosis -2.00809e-01
conditions:
norm T=300 s, I=19.9 muA
pol0 1 raw
the Standard Model, rst paperback edition (with correc-
tions) (Cambridge University Press, Cambridge, 1992).
Expected elastic: 2.72263e+05 Expected elastic: 2.48168e+05 Expected elastic: 2.24327e+05
norm pol1 1 dnl norm pol1 1 dnl norm pol1 1 dnl
4000
6000
4000
6000
4000
11. M.M. Pavan et al., PiN Newslett. 16, 110 (2002).
2000 2000 2000 12. X.-D. Ji, Phys. Rev. Lett. 74, 1071 (1995).
0
0 20 40 60 80 100
0
0 20 40 60 80 100
0
0 20 40 60 80 100 13. M. Goncharov et al., Phys. Rev. D 64, 112006 (2001).
14. M. Tzanov et al., Nutev neutrino dis., hep-ex/0306035,
Fig. 12. The blue histograms show electron spectra where 2003.
the additional scintillator system had given a coincidence sig- 15. M. Gluck, E. Reya, A. Vogt. Eur. Phys. J. C 5, 461 (1998).
nal. They exhibit a clear elastic peak with a good signal-to- 16. E. Leader et al., Phys. Rev. D 67, 074017 (2003).
background ratio. The black and red histograms (only partly 17. D.B. Kaplan et al., Nucl. Phys. B 310, 527 (1988).
shown) correspond to the large background from photons from 18. M.J. Musolf et al., Phys. Rep. 239, 1 (1994).
0 -decay. Both spectra are taken always in parallel for better 19. J. Friedrich, Th. Walcher, Eur. Phys. J. A 17, 607 (2003).
control of systematic eects. 20. W.J. Marciano, A. Sirlin, Phys. Rev. D 29, 75 (1984).
F.E. Maas: Parity-violating electron scattering at the MAMI facility in Mainz (A4) 115
21. S. Eidelmann et al., Review of particle properties, Phys. 28. P. Achenbach et al., Nucl. Instrum. Methods A 465, 318
Lett. B 592, 1 (2004). (2001).
22. S.-L. Zhu et al., Phys. Rev. D 62, 033008 (2000). 29. F.E. Maas et al., Phys. Rev. Lett. 93, 022002 (2004).
23. G. Kuster, H. Arenhovel, Nucl. Phys. A 626, 911 (1997). 30. F.E. Maas et al., Evidence for strange quark contributions
24. K. Aulenbacher et al., Nucl. Instrum. Methods A 391, 498 to the nucleons form-factors at q 2 = 0.108 (Gev/c)2 , nucl-
(1997). ex/0412030, 2004.
25. M. Seidl et al., High precision beam energy stabilisation of 31. D.T. Spayde et al., Phys. Lett. B 583, 79 (2004).
the Mainz microtron MAMI, in Proceedings of the EPAC 32. D.B. Leinweber et al., Phys. Rev. Lett. 94, 212001 (2005).
2000 (2000) p. 1930. 33. R. Lewis et al., Phys. Rev. D 67, 013003 (2003).
26. P. Bartsch, Aufbau eines Meller-Polarimeters fur 34. T.R. Hemmert et al., Phys. Rev. C 60, 045501 (1999).
die Drei-Spektrometer-Anlage und Messung der He- 35. A. Silva et al., Eur. Phys. J. A 22, 481 (2004).
lizitatsasymmetrie in der Reaktion p(e,e p) 0 im Bereich 36. V. Lyubovitskij et al., Phys. Rev. C 66, 055204 (2002).
der -Resonanz, Dissertation Mainz, 2001. 37. H. Weigel et al., Phys. Lett. B 353, 20 (1995).
27. F.E. Maas et al., Proceedings of the ICATPP-7 (World 38. H.-W. Hammer et al., Phys. Rev. C 60, 045204 (1999).
Scientic, 2002) p. 758.
Eur. Phys. J. A 28, s01, 117 127 (2006)
DOI: 10.1140/epja/i2006-09-013-6 EPJ A direct
electronic only
/
Published online: 17 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Virtual Compton Scattering (VCS) o the proton is a recent eld of investigation of the nucleon
structure. VCS at threshold gives access to the Generalized Polarizabilities (GPs) of the proton. The
qualities of both the beam and the high-resolution spectrometers available at the Mainz Microtron MAMI
allowed us to perform at rst such delicate experiments. This paper deals with dierent experiments
dedicated to the GPs measurements. They are realized without and with polarization, below and just
above pion threshold.
PACS. 13.60.Fz Elastic and Compton scattering 14.20.Dh Protons and neutrons 25.30.Rw Electro-
production reactions
1 Virtual Compton Scattering and the world global average of the electric (E ) and magnetic
polarizabilities (M ) polarizabilities on the proton is based on an experi-
mental study investigated at MAMI with the tagged real
One of the main challenges of hadronic physics in the photon beam [1].
regime of strong (non-perturbative) QCD is to identify E =(11.9 0.5(stat.) 1.3(syst.) 0.3(mod.))104 fm3 ,
the relevant degrees of freedom of the nucleon. Though
the small distance structure is rather well described by M =(1.2 0.7(stat.) 0.3(syst.) 0.4(mod.))104 fm3 .
point-like quarks and gluons, its structure at larger dis- We can note the small size of the polarizabilities which
tance is not so well understood. There exist many models reveals the feature that the nucleon is strongly bound.
ranging from constituent quark models to chiral models. For comparison the electric polarizability of the hydrogen
Polarizabilities are one of the fundamental observ- atom is of the order of the atomic volume, and the electric
ables to describe the internal structure of the nucleon and polarizability of the proton E is only 0.05 per cent of its
they have been investigated with real Compton scattering volume. Furthermore the magnetic polarizability is still
(RCS) since the early 1950s. As the light scattering on smaller, one tenth of the electric polarizability.
atmospheric atoms which gives the well known Rayleigh Virtual Compton scattering (VCS) o the proton
eect for blue skies and red sunsets through oscillation refers to the reaction p p, where stands for an
of the electrons inside the atoms, real Compton scatter- incoming virtual photon of four-momentum squared Q2 .
ing sheds light on the nucleon structure. This is clearly This reaction is experimentally accessed through photon
illustrated in a common denition of the electric polariz- electroproduction ep ep. The corresponding Feynman
ability E in a non relativistic approach at the rst-order diagram is indicated in g. 1.
perturbation for an applied electric dipole moment D: In the 1960s the VCS appeared as a rather un-
wanted contribution to radiative corrections to electron
|N |Dz |N |2
E = 2N =
N . scattering on a proton [2]. It was mentioned as proton
EN E N Bremsstrahlung. In 1974 Arenhovel and Drechsel [3], from
In this formula N indicates each nucleon resonance. The the Institut fur Kernphysik at Mainz, considered the VCS
polarizability is then sensitive to all the excitation spec- for the rst time as a good way to measure generalized
trum of the nucleon (even if the low energy of the per- polarizabilities (GPs). Only in 1995 with the new genera-
turbating photon does not allow the real formation of the tion of facilities of high duty cycle to investigate exclusive
nucleon resonances). reactions, and with new theoretical concepts it regained
The Mainz laboratory has a long tradition in this eld. interest.
Several experiments have been dedicated to the determi- The general theoretical framework for VCS at thresh-
nation of proton, neutron or pion polarizabilities. Today old has been extensively described by Guichon et al. [4,5]
and the Mainz theoretical group conducted by Drechsel [6,
a
e-mail: ndhose@cea.fr 7]. VCS reaction at threshold means that the produced
118 The European Physical Journal A
Fig. 1. The VCS graph for the proton. Non relativistic constituent quark model (CQM) is
based on the assumption that baryons are composed of
three massive quarks moving within a harmonic oscil-
photon has a small enough momentum or that its electric lator conning potential and additional hyperne in-
(E) and magnetic (M) elds look constant over the size of teractions. One of its success is to explain most of the
the nucleon. In the following the three-momenta absolute observed nucleon resonance mass spectrum. Calcula-
values of the virtual and real photons in the photon-proton tions have been performed in this framework by Gui-
center of mass (c.m.) system are noted q and q , respec- chon, Liu and Thomas [4,9] and Pasquini, Scherer and
tively, and vary independently, this is in contrast with real Drechsel [10].
Compton scattering where q = q . Here the low momen- Phenomenological approach can be realized with an
tum q of the produced real photon denes the size of the eective Lagrangian model (ELM). Such a calculation
electromagnetic (EM) perturbation, while the momentum has been performed by Vanderhaeghen [11] which in-
of the virtual photon q (or the four-momentum squared cludes the eects of all the rst nucleon resonances and
Q2 ) sets the scale of the observation of the nucleon inter- 0 exchange in the t channel.
nal structure.
These two rst kinds of model describe well all the reso-
In the low momentum regime the reaction can be in-
nant contributions, but not the non-resonant one. Their
terpreted as electron scattering on a nucleon placed in a
limitation is that they have no relationship to chiral sym-
quasi-constant applied EM eld [5]. The induced motion
metry. This is an important property of QCD which gov-
of the nucleon as a whole can be eliminated thanks to a
erns much of low-energy hadron physics. The pion is the
low-energy theorem [8], so one is left with the deformation,
Goldstone boson of spontaneously broken chiral symme-
due to the applied eld, of the nucleon internal currents
try, and plays a very special and major role at low energy.
J (r) and the electron scattering measures its Fourier
The two next groups of calculations respect chiral sym-
transform J (Q). To lowest order in QED , J (Q) is
metry.
linear in the applied eld and the 6 coe cients of propor-
tionality are the GPs [4,5,6,7]. When Q2 = 0 two of them A simple model to describe interaction of Dirac parti-
reduce to the usual polarizabilities E and M measured cles with a chiral eld is the linear sigma model (LSM)
in real Compton scattering. Analogously to the form fac- in the limit of an innite sigma mass. Though this
tors for elastic scattering, which describes the charge and model is not a very realistic description of the nucleon,
magnetization distributions, VCS gives access to the de- nevertheless it fullls all the relevant symmetries like
formation of these distributions by an external EM eld, Lorentz, gauge and chiral invariance. A complete cal-
and will yield valuable information about the non pertur- culation of all the one-loop diagram contributions (for
bative structure of the nucleon. the photon interaction with a nucleon-pion system) has
This can be illustrated by a very naive picture of the been performed by Metz and Drechsel [12].
polarizabilities which are the results of an electromag- Chiral perturbation theory (ChPT) is a very system-
netic perturbation applied to the nucleon components. An atic and consistent approach with a most general La-
electric eld moves positive and negative charges inside grangian based on QCD symmetries. Heavy-baryon
the proton in opposite directions. The induced electric chiral perturbation theory allows for a systematic per-
dipole moment is proportional to the electric eld, and turbative expansion in powers of small parameters
the proportionality coe cient is the electric polarizability (no- ted p) as quark masses, inverse of hadron masses
E which measures the rigidity of the proton. A magnetic or external momenta. Hemmert, Holstein, Knocklein
eld acts dierently on the quarks and the pion cloud. The and Scherer [13] have performed a third order O(p3 )
quarks (of spin 1/2) align their magnetic moment parallel calculation for all the GPs while Kao and Vander-
to the magnetic eld giving the strong magnetic excitation heaghen [16] have performed a fourth order O(p4 ) cal-
of the (1232) resonance. The pions are at low energy, an culation but only for the spin polarizabilities which
essential element of the structure of the nucleon notably at exclude predictions for E and M . Nevertheless pre-
its surface giving the famous representation of a pion cloud vious calculations for E (Q2 ) and M (Q2 ) have been
surrounding the nucleon. The pions (of spin 0) distributed realized at Q2 = 0 by Bernard, Kaiser, Schmidt
at the surface of the proton, will generate eddy currents. and Meissner [17] including all terms to order O(p4 )
N. dHose: Virtual Compton Scattering at MAMI 119
M (10 fm )
3
12 5 p p p p p p
-4
-4
10 4
3 (a) (b) (c)
8
2 Fig. 3. The p(e, e p) reaction. The initial, nal electron and
6
1 initial, nal proton quadri-momenta are k, k and p, p re-
4 0 spectively. The nal photon quadri-momentum is q . In the
2 -1 one photon exchange approximation, a) and b) correspond to
0 -2 the Bethe-Heitler (BH) process. c) corresponds to the Vir-
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 tual Compton Scattering (VCS) process. We note q the quadri-
2 2 2 2
Q (GeV ) Q (GeV ) momentum of the virtual photon exchanged in the VCS pro-
cess, that is q = k k (Q2 = q 2 ).
Fig. 2. Evolution of the electric and magnetic polarizabili-
ties with Q2 . Experimental results [1] at Q2 = 0 (with only
statistical errors) and ve theoretical predictions CQM [10], complementary kinematical regimes at Jeerson Lab.
ELM [11], LSM [12], ChPT [13] and DR [15] are reported. See (Q2 = 1 and 2 GeV2 ) [19] in 1998 and MIT-Bates (Q2 =
the text for comments. 0.05 GeV2 ) [20] in 2000. They are long and delicate exper-
iments and they rely on a careful analysis of the data.
for a phase space factor. M is the coherent sum of the Table 2. List of the 10 GPs with the corresponding electro-
dierent amplitudes: magnetic transitions. Their relation with the polarizabilities
1 BH 2 obtained in real Compton scattering are indicated. 6 GPs are
M= T + T V CS , (2) independent. Our choice is a priori arbitrary, and is realized
4 spin
by the 6 surrounded GPs.
1 BH 2
= T + T Born + T N onBorn . EM transition VCS GPs RCS polarizabilities
4 spin Q2 0
M(C0 M 1)S=1
P (11,00)1 0
The low energy theorem (LET) from Low [8] states that
in an expansion in powers of the real photon energy q M (C2 M 1)S=1
P (11,02)1
8 4
27 e2
(2 + 4 )
(but xed arbitrary q), the rst term of the amplitudes M (M 1 M 1)S=0
P (11,11)0
8 4
(M )
1 3 e2
T BH and T Born is of the order q (well-known infra-
M (M 1 M 1)S=1
P (11,11)1
0
red divergence), while the rst term of T N onBorn is of the
M(C2,E2 M 1)S=1
P (11,2)1 = 0
order q :
1
0.7
0.6
0.5
0.4
0.3
0.2
-100 0 -100 0
0.2
0.1
0.09
0.08
0.07
0.06
-100 0 -100 0 -100 0
(deg)
1600
determined by the acceptance of the two spectrometers
around 0 and 180 . The spherical angles and are de-
1400 ned such that = 0 corresponds to the half plane con-
taining the electron momenta. To ease the presentation
1200 the data are plotted with ranging from 180 to +180 ;
1000
the negative values corresponding in fact to = 180 .
The wide range of from 141 to +6 covers the back-
800 ward direction relative to the incoming and outgoing elec-
trons. Here, the VCS contributions are dominant because
600 the electron radiations (BH) are emitted predominantly
in the electron directions.
400
The cross sections d5 BH+Born are presented by
200 the solid lines in g. 6. At small photon momentum
q = 33.6 MeV/c the agreement between the radiatively
0
corrected data and d5 BH+Born is excellent, and the
-5000 0 5000 10000 15000 20000 25000
2 2
Mx (MeV ) deviation from this known cross section increases when
q increases, as expected from the eect of the proton
Fig. 5. Missing mass spectrum obtained for the setup at q = polarizabilities. In order to determine accurately the
111.5 MeV/c. polarizabilities, a careful analysis of possible systematic
errors on the deviation is of particular importance. First
the BH and Born contributions rely on the knowledge of
around zero, which was possible thanks to the excellent the proton form factors. Consequently we also measured
resolution of the facility (momentum resolution of 104 the absolute elastic scattering cross section for each
and angular resolution better than 3 mrad) (see g. 5). kinematic setting of the VCS experiment. These measure-
The aim of this rst VCS experiment below pion ments validate the use of the form factor parametrization
threshold was to measure the ve-fold dierential cross from Hohler [22] at a precision better than 1%. Second
sections in a wide photon angular range, at 5 values of the the radiative corrections, which are of the order of 20% of
photon momentum q : 33.6, 45, 67.5, 90, and 111.5 MeV/c the cross section, have been evaluated by Vanderhaeghen
(presentation in g. 6). The 3 other kinematical variables et al. [23]. The systematic uncertainties are estimated
were held xed, namely the virtual photon momentum, to equal 2% for the calculation performed to order 4
q = 600 MeV/c (Q2 = 0.33 GeV2 ), the virtual photon in the VCS cross section. Third the luminosity and the
polarization = 0.62. The out-of-plane angle range is detector e ciencies are controlled within an accuracy
122 The European Physical Journal A
5 5 BH+Bo , , -2
(d -d )/q = 0 + q 1 + ... (in GeV )
0.5 10
0 /v2 (GeV-2)
0.25 7.5
0
5
-0.25
0.5 -0.5 2.5
0.25 0
0
-2.5
-0.25
0.5
-0.5 -5
0.25 -7.5
0
-10
-0.25
0.5
-0.5 -0.1 0 0.1 0.2 0.3 0.4 0.5
v1/v2 ()
0.25
0 Fig. 8. Compilation of the complete data set (for the 14 angles)
of 0 /v2 as a function of v1 /v2 . The data are reasonably well
-0.25 aligned; the errors indicated are statistical only. This allows
-0.5 to extract the two structure functions PLL PTT / and PLT
with statistical errors and the 2 given.
,
q (MeV/c)
Fig. 7. (d5 d5 BH+Born )/q studied as a function of the Figure 8 presents the complete data set (for the 14
real photon momentum q for the 14 measured scattering an- angles) of 0 /v2 as a function of v1 /v2 (cf. eq. (6)). The
gles . The intercept at origin is 0 . In a rst method it is data are reasonably well aligned, which suggests that the
determined at each scattering angle by the mean value in higher-order terms in the expansion of the cross section
the investigated real photon momentum range (solid line). The (cf. eq. (5)) are not so important. This good alignment for
dash-dotted, dashed and dotted lines show evolutions in the a wide angular range is also indicative of the consistency of
framework of the 2nd, 3rd and 4th methods, respectively. the experimental data. We extract the two structure func-
tions PLL PTT / and PLT as the slope and intercept
of a linear t to the data (according to eq. (6)) [18].
of 1%, the solid angles are determined within an Second method: We make the hypothesis of a linear
accuracy of 2% using a Monte Carlo [24] simulation evolution with the real photon momentum for each angle
which reproduces perfectly the missing mass spectra. All which is tted to the data. The result is indicated by the
these uncertainties are constant over the angular range dash-dotted line in g. 7.
of the real photon and are controlled by the fairly good Third method: The q evolution is supposed to be gov-
agreement between the radiatively corrected data and the erned by the interference between the complete BH and
predicted BH and Born cross section at small q . However Born amplitudes considered at all order in the q expan-
small imperfections in the spectrometer optic calibration sion (complete eq. (3)) and the Non-Born amplitude trun-
which could provide distortion of the angular distributions cated at the rst order in the q expansion (truncated
are estimated to give a variation of cross section of 2.5%. eq. (4)). The only parameters are a priori the 6 gen-
Figure 7 shows the behavior of (d5 d5 BH+Born )/ eralized polarizabilities contained in the rst and only
q as a function of the real photon momentum q for the
term of the considered Non-Born amplitude. They are
14 measured scattering angles . The goal is to determine adjusted with a best t on the complete set of 14 5
the intercept at origin (noted 0 in eq. (5)), and this gure data. In order to have a better convergence, the polariz-
illustrates the basic di culty of this experiment that is the ability P (01,01)0 (Q2 = 0.33 GeV2 ) is xed by the result
increase of the statistical errors when q decreases. Four obtained in real Compton scattering scaled by the electric
methods are then considered in the following. form factor and P (11,02)1 (Q2 = 0.33 GeV2 ) is xed at 0 (it
First method based on the LET: As is apparent in g. 7, corresponds to the quadrupolar deformation of the N-
there is no strong evolution with the real photon momen- transition which is expected to be very small). The result
tum. Therefore we make the hypothesis that there is no of the t for the 4 remaining polarizabilities is presented
q dependence in (d5 d5 BH+Born )/q . 0 is then de- in table 3. This third method, mainly realized to jus-
termined at each scattering angle by the mean value of tify a rather at q evolution of (d5 d5 BH+Born )/q
the data at the 5 photon energies. presented by the dashed points in g. 7, allows one to
N. dHose: Virtual Compton Scattering at MAMI 123
Table 3. Results for the polarizabilities extracted in the third Table 4. The structure functions determined in the MAMI
method. These results are compared to the heavy-baryon chiral experiment using the four methods and compared to model
perturbation theory (HBChPT) predictions [13]. predictions at Q2 = 0.33 GeV2 and = 0.62. The errors are
statistical only, except for the rst method where two system-
Third Method HBChPT Units
atic errors are indicated in brackets. The prediction for DR
PLL P (01,01)0 0.0626 xed 0.056 fm3 model is given for 2 values of and close to the values
PTT P (11,11)1 +0.0048 0.0034 +0.001 fm3 determined experimentally.
P (01,12)1 0.0123 0.0026 0.008 fm4
PLT P (11,11)0 0.0384 0.0186 0.034 fm3 Q2 = 0.33 GeV2 PLL PTT / PLT
P (11,00)1 0.157 0.070 0.096 fm2 = 0.62 (GeV 2 ) (GeV 2 )
method 1 [18] 23.7 2.2 5.0 .8
P (11,02)0 0. xed +0.003 fm4
(2 = 1.4) (4.3 0.6) (1.4 1.1)
method 2 23.7 8.1 7.8 3.0
(2 = 1.3)
determine some spin polarizabilities with reasonable pre- method 3 33.6 11.7 6.5 4.2
cision (notably P (01,12)1 ). (2 = 1.7)
method 4 23.2 3.0 3.2 2.0
Fourth method using Dispersion Relations: This
(2 = 1.5) ( = 1.6 0.2) ( = 0.5 0.2)
method was used after the publication [18] of the rst
HBChPT [13] 26.3 5.7
VCS MAMI experiment at the sight of the other exper-
LSM [12] 10.9 0.2
iments of JLab and MAMI where a rather at q evo-
ELM [11] 5.9 1.9
lution was not so obviously conrmed by the data. This
NRQCM [9] 11.0 3.5
method is based on the formalism of Dispersion Relations
NRQCM [10] 14.7 4.5
(DR) [14,15] for the invariant VCS amplitudes and works
DR [15] 22.0 5.5
below pion threshold as well as in the rst resonance re- ( = 1.4 GeV) ( = 0.5 GeV)
gion. Assuming analyticity, crossing symmetry and an ap-
propriate high-energy behavior, unsubtracted dispersion
relations relate the real part of VCS amplitudes to an in- a dipole parametrization has also been proposed:
tegral over the virtual photon energy of a function of their
imaginary part. The imaginary part of a VCS amplitude (E E
N
)Q2 =0
is given by the sum of N intermediate states, computed E (Q2 ) E
N
(Q2 ) = (9)
(1 + Q /2 )2
2
from N N data (in the phenomenological MAID-
2000 analysis [25]), plus higher order contributions beyond The mass scale is the second free parameter of the DR
N . Moreover asymptotic contributions have also to be formalism which can be extracted from a t to the VCS
considered for two VCS amplitudes (F1 and F5 ) which data at dierent Q2 values.
cannot fulll unsubtracted dispersion relation framework. The evolution with the real photon momentum q of
The t-channel 0 exchange and the knowledge of the the MAMI VCS data obtained at Q2 = 0.33 GeV2 is rela-
F0 form factor x the asymptotic contribution to F5 tively sensitive to the choice of the free parameter values:
and determine completely the spin-dependent GPs. = 1.6 0.2 GeV and = 0.5 0.2 GeV. The cor-
The asymptotic contribution of the amplitude F1 related responding evolution with the real photon momentum q
to the polarizability P (11,11)0 or M (Q2 ) originates from is presented by the dotted line in g 7. This prediction is
the t-channel intermediate states. In a phenomenolog- rather close to the evolution of method 3 given by the in-
ical analysis, this continuum is parametrized through the terference between the complete BH+Born amplitude and
exchange of a scalar-isoscalar particle in the t-channel, the truncated Non-Born amplitude, except for close to
i.e. an eective -meson which gives rise to a diamag- 0 . This kinematical point is only sensitive to PLT and
netic contribution. The asymptotic part and the disper- this indicates dierent results for this observable in the
sive contributions beyond N are estimated using a dipole framework of these 2 methods.
parametrization of the dierence: Results: Table 4 presents the two structure functions
in the framework of the four methods and compares them
(M MN
)Q2 =0 to theoretical predictions presented in the introduction:
M (Q2 ) M
N
(Q2 ) = (8) the heavy-baryon chiral perturbation theory calculation
(1 + Q /2 )2
2
(HBChPT) [13], the linear sigma model (LSM) [12], the ef-
fective Lagrangian model (ELM) [11], two non-relativistic
The mass scale is a free parameter related to the dia- constituent quark models (NRQCM) [9,10] and the dis-
magnetism distribution inside the nucleon. It can be ex- persive relation approach (DR) [15]. The three errors for
tracted from a t to the VCS data at dierent Q2 values. the rst method are, respectively, statistical error on the
Though unsubtracted dispersion relation is valid for data, systematic error on photon angular distributions,
the amplitude F2 related to the polarizability P (01,01)0 and systematic error on the normalization. For the other
or E (Q2 ), it is particularly relevant to wonder about the methods only statistical errors are reported. The predic-
quality of the saturation of the subtracted dispersion in- tion for the DR model is given for 2 values of and
tegrals by N intermediate states only. For this purpose close to the values determined experimentally.
124 The European Physical Journal A
60 60
2.3 The two other unpolarized experiments at JLab
40 40 and MIT-Bates
20 20 The JLab experiment E93-050 [19] has been performed
in the Hall A of the Thomas Jeerson National Accel-
0 0 erator Facility at Q2 = 0.9 and 1.8 GeV2 . The values of
0 0.2 0.4 0 0.2 0.4
Q (GeV2)
2
Q2 (GeV2) are 0.95 and 0.88, respectively. Data cover the region
below the pion threshold and the resonance region up to
s = 2 GeV. The experimental analysis of the complete
2 2
experiment is presented in ref. [19].
0 0
The Bates experiment 97-03 [20] has been performed at
2 2 Q2 = 0.05 GeV2 and = 0.90. Measurements have been
4 4 done in-plane and at 90 out-of-plane, using the OOPS
6 6 spectrometers. The experiment covers a limited range in
8 8 polar angle around 90 , so the structure functions are
10 10 determined from the -dependence of the cross section.
12 12 Data analysis is still in progress and only preliminary re-
0 0.2 0.4 0 0.2 0.4 sults [20] can be presented. This experiment represents a
Q2 (GeV2) Q2 (GeV2) laboratory achievement, having made the rst use of the
high duty factor beam in the South Hall Ring and of the
Fig. 9. Comparison of the unpolarized structure functions de-
full OOPS system.
termined in the VCS MAMI experiment at Q2 = 0.33 GeV2
and in the RCS results [1] with the predictions of the DR for-
malism [15] (left panel) and of the O(p3 ) HBChPT [13] (right
panel). The upper panels give the result for PLL PTT / and
the lower panel for PLT . The contributions of the scalar GPs
are indicated by the dashed (or dotted) lines and the total con- 10 4
tributions of the scalar and spin-dependent GPs are indicated
by the solid (or dashed-dotted) lines. The DR prediction for
the scalar GP E (Q2 ) is calculated for = 1.4 GeV (upper
left panel). Two values of are used to calculated the scalar
2
GP M (Q2 ) contribution (lower left panel) in order to show
the sensitivity. The contributions for = 0.4 and 0.6 GeV 5
are used in the dotted and dashed lines, respectively. Figure
extracted from [15].
0
d 5 (nb/GeV sr 2)
0.9
0.8
2 0.7
0.6
0 0.5
0.4
-2
0.3
-4
-6 0.2
-8
0.05 0.1 0.15 0.2 0.25 0.3
0 0.2 0.4 ,
2 2 q (GeV)
Q (GeV )
Fig. 12. VCS dierential cross section as a function of the real
Fig. 11. PTT evaluated using the DR model [15] (solid
photon energy q in the MAMI kinematics below and above
line), the O(p3 ) HBChPT [13] (dotted line) and the O(p4 )
the pion threshold. The VCS MAMI data are reported. The
HBChPT [16] (dashed line).
BH+Born contribution is given by the dash-dotted line. Pre-
dictions for the total cross section are given in the DR ap-
proach [15] using a xed value of = 1 GeV and for three
We present in g. 10 the world results for the elec- values of : 0.6 GeV (solid line); 0.7 GeV (dotted line) and
tric E (Q2 ) and magnetic M (Q2 ) GPs deduced from the 0.4 GeV (dashed line). Fig extracted from ref. [15].
MAMI, JLab and MIT-Bates experiments. The value of
the electric E (Q2 ) and magnetic M (Q2 ) GPs can de-
termined directly by the coe cients and obtained 3 Single polarized experiments above pion
in the DR analysis and using eq. (9) and eq. (8) or indi- threshold
rectly by the structure functions PLL PTT / and PLT
determined in the LET analysis in which the spin GPs con- Figure 12 shows the DR predictions for photon energies
tributions are evaluated in the DR model and subtracted ranging from threshold to the (1232)-resonance region.
using eq. (7). The agreement between these 2 methods The deviation from the BH+Born prediction rises strongly
was reasonably controlled in the JLab experiment (see the after pion threshold. When crossing the pion threshold,
dierent results at Q2 = 0.9 and 1.8 GeV). The direct de- the VCS amplitude acquires an imaginary part due to
termination allows us to use also the VCS data in the res- the coupling to the N channel. Therefore single polariza-
onance region (see in g. 10 the result at Q2 = 0.9 GeV2 tion observables become non-zero above pion threshold. A
with the smallest statistical error.) particularly relevant observable is the electron single spin
The curves in g. 10 are calculated using the DR model asymmetry (SSA) which is obtained by ipping the elec-
and the dierent values of and obtained in each ex- tron beam helicity. For VCS this observable is mainly due
periment. By denition all the DR predictions (see eqs. (9) to the interference of the real BH+VCS amplitude with
and (8)) are constrained to go through the experiment the imaginary part of the VCS amplitude. As the SSA van-
RCS point at Q2 = 0. The fact that there is no unique ishes in-plane, its measurement requires an out-of-plane
DR curve going through all the data points, especially for experiment. Such an experiment has been proposed at
the electric polarizability, does not invalidate the model. It MIT-Bates [27] and is being realized at MAMI [28] thanks
simply means that the dipole parametrization of eqs. (9) to one of the spectrometers of the A1 collaboration moving
and (8) does not hold over the entire Q2 range. Another out-of-plane. In g. 13, the SSA is presentedfor a kinemat-
fact to be aware of is the model-dependency introduced ics in the (1232) region, corresponding to s = 1.2 GeV.
in this gure by transforming the structure functions into The DR calculation shows that the SSA is quite sizeable,
GPs. The spin-dependent GPs are evaluated using the DR and it is mainly sensitive to the imaginary part of the VCS
model, and as it has been pointed in g. 9, this evaluation amplitude, displaying only a rather weak dependence on
is quite smaller than in the O(p3 ) HBChPT [13]. the GPs (obtained for the dierent values of and ).
Therefore it provides an excellent cross-check of the dis-
It is clear that measurements of individual scalar and persive input (MAID 2000) in the DR formalism for VCS,
spin-dependent GPs are necessary to go further. We can in particular by comparing at the same time the pion and
note in g. 11 the very dierent predictions for PTT using photon electroproduction channels through the excita-
the DR model [15] or the O(p3 ) HBChPT [13] or else tion. The MAMI analysis is still in progress.
the O(p4 ) HBChPT [16]. An extraction of PTT can be
achieved in a further experiment at MAMI at the same
Q2 = 0.33 GeV2 , but with an other value of . We can 4 Double polarized experiments at threshold
take the benet of the next 1.5 GeV energy of the beam
to access a new value of to have a comfortable lever arm A double-polarization VCS experiment is also presently
for a longitudinal-transverse separation. being realized at MAMI. The theoretical framework of
126 The European Physical Journal A
0 eYcm
e
*
eXcm
0.05
e p
p eZcm
0.1
0.15
0 10 20 30 40 50 60 70 80 90
moving out-of-plane. It is clearly a very challenging exper- 6. D. Drechsel, G. Knochlein, A. Metz, S. Scherer, Phys. Rev.
iment, relying on a very delicate expertise of the complete C 55, 424 (1997).
apparatus and requiring high statistics and very reduced 7. D. Drechsel, G. Knochlein, A. Yu Korchin, A. Metz, S.
systematic errors. Scherer, Phys. Rev. C 57, 941 (1998).
8. F.E. Low, Phys. Rev. 110, 974 (1958).
9. G.Q. Liu, A.W. Thomas, P.A.M. Guichon, Austral J. Phys.
49, 905 (1996).
5 Conclusion 10. B. Pasquini, S. Scherer, D. Drechsel, Phys. Rev. C 63,
025205 (2001).
An ambitious program to reach the generalized polariz- 11. M. Vanderhaeghen, Phys. Lett. B 368, 13 (1996).
abilies of the proton has been undertaken at MAMI over 12. A. Metz, D. Drechsel, Z. Phys. A 356, 351 (1996); 359,
the last ten years. The ultimate Grail is the separation 165 (1997).
between spin-independent and spin-dependent GPs which 13. T.R. Hemmert, B.R. Holstein, G. Knochlein, S. Scherer,
seems very promising for the study of the nucleon struc- Phys. Rev. D 55, 2630 (1997); Phys. Rev. Lett. 79, 22
ture. (1997); T.R. Hemmert, B.R. Holstein, G. Knochlein, D.
All the results reported here are the fruits of the Drechsel, Phys. Rev. D 62, 014013 (2000).
complete A1-VCS collaboration. I would like to acknowl- 14. B. Pasquini, M. Gorchtein, A. Metz, M. Vanderhaeghen,
edge all the students, Luca Doria, Peter Janssens, Imad Eur. Phys. J. A 11, 185 (2001).
Bensafa, Jan Friedrich, Julie Roche, David Lhuillier, Do- 15. D. Drechsel, B. Pasquini, M. Vanderhaeghen, Phys. Rep.
minique Marchand for which the work was essential to pro- 378, 99 (2003).
duce reliable results on these very meticulous experiments. 16. C.W. Kao, M. Vanderhaeghen, Phys. Rev. Lett. 89, 272002
I wish to underline the strong support and the synergy (2002).
given by Helene Fonvieille, Harald Merkel, Michael Dis- 17. V. Bernard, N. Kaiser, A. Schmidt, U. Meissner, Phys.
tler, Luc Van Hoorebeke, Gabriel Tamas, Robert Van de Lett. B 319, 269 (1993); Z. Phys. A 348, 317 (1994).
Vyver, Jorg Friedrich, Thomas Walcher for this research. 18. J. Roche et al., Phys. Rev. Lett. 85, 708 (2000).
19. G. Laveissiere et al., Phys. Rev. Lett. 93, 122001 (2004).
I am also very grateful to Dieter Drechsel, Pierre Guichon,
20. J. Shaw, R. Miskimen, MIT-Bates Proposal 97-03, (1997)
Marc Vanderhaeghen, Barbara Pasquini, Stefan Scherer,
and P. Bourgeois, PhD Thesis.
Thomas Hemmert, Ulf Meissner for their pedagogical lec- 21. K.I. Blomqvist et al., Nucl. Instrum. Methods A 403, 263
tures and theoretical support in the data interpretation. (1998).
It is clear that the success of the VCS MAMI exper- 22. G. Hohler, E. Pietarinen, I. Sabba-Stefanescu, F.
iments has its origin in the coherent eort between ex- Borkowski, G.G. Simon, V.H. Walther, R.D. Wendling,
cellent physicists as Karl-Heinz Kaiser, always concerned Nucl. Phys. B 114, 505 (1976); private communication.
with the performance and the high quality of the electron 23. M. Vanderhaeghen, J.M. Friedrich, D. Lhuillier, D. Marc-
facility, Thomas Walcher strongly supporting and man- hand, L. Van Hoorebeke, J. Van de Wiele, Phys. Rev. C
aging all the eorts for such a challenging experimental 62, 025501 (2000).
program, Jorg Friedrich, expert of delicate and precise ex- 24. P. Janssens, L. Van Hoorebeke et al., to be published in
periments and Dieter Drechsel, stimulating progress in our Nucl. Instrum. Methods.
scientic knowledge. 25. D. Drechsel, O. Hanstein, S.S. Kamalov, L. Tiator, Nucl.
Phys. A 645, 145 (1999).
26. H. Fonvieille, Proceedings of the Erice School of Nuclear
Physics, 26th Course, Prog. Part. Nucl. Phys. 55, 198
References (2005) and private communication.
27. N.I. Kaloskamis, C.N. Papanicolas, MIT-Bates proposal
1. V. Olmos de Leon et al., Eur. Phys. J. A 10, 207 (2001). (1997).
2. Y.S. Tsai, Phys. Rev. 122, 1898 (1961). 28. N. dHose, H. Merkel, MAMI Proposal (2001).
3. H. Arenhovel, D. Drechsel, Nucl. Phys. A 233, 153 (1974). 29. M. Vanderhaeghen, Phys. Lett. B 402, 243 (1997).
4. P.A.M. Guichon, G.Q. Liu, A.W. Thomas, Nucl. Phys. A 30. Th. Pospischil et al., Nucl. Instrum. Methods A 483, 726
591, 606 (1995). (2002).
5. P.A.M. Guichon, M. Vanderhaeghen, Prog. Part. Nucl. 31. C.W. Kao, B. Pasquini, M. Vanderhaeghen, Phys. Rev. D
Phys. 41, 125 (1998). 70, 114004 (2004).
Eur. Phys. J. A 28, s01, 129 137 (2006)
DOI: 10.1140/epja/i2006-09-014-5 EPJ A direct
electronic only
/
Published online: 15 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Over the last decade, a series of dedicated experiments to test heavy baryon chiral perturbation
theory was performed at MAMI. Photo production of neutral pions close to threshold with unpolarized and
polarized photon beam was performed to separate the multipole amplitudes at threshold. The extension of
this experiments to a modest photon virtuality of Q2 < 0.1 GeV2 /c2 was performed to extract additional
longitudinal multipoles and to exploit the Q2 evolution predicted by theory. An out-of-plane measurement
above + threshold with polarized electron beam gave access to the imaginary part of the s-wave amplitude.
Finally, by coherent photo and electro production from the deuteron the neutron amplitude could be
extracted.
PACS. 25.30.Rw Electroproduction reactions 13.60.Le Meson production 12.39.Fe Chiral Lagrangians
2.5 from SAL [6] and MAMI [7]. The data were taken at
the tagged photon beam of the MAMI A2 Collaboration
2.0 (g. 2). The reaction was identied by the detection of
1.5 the two decay photons of the pion in coincidence with the
TAPS detector, an array of 504 BaF detector modules ar-
1.0 ranged in 7 blocks.
0.5 With this setup, dierential cross sections were mea-
sured up to an incoming photon laboratory energy of
0.0
144 147 150 153 156 159 162 165 168 168 MeV. By tting the angular coe cients A, B, and C to
E / MeV the dierential cross section and extrapolating to thresh-
old by using the known energy dependence, the multipole
Fig. 1. Total cross section of neutral pion photo production combinations Re E0+ , P1 , and P22 +P32 could be extracted.
(from [3]).
= 0 (1 P () cos 2)
e am
with the angle between polarization plane and produc-
B
tion plane and P the photon polarization. This asymme-
try is proportional to the dierence P22 P32 and allows
the decomposition of the modulus of these two multipole
amplitudes.
At MAMI, a polarized photon beam was prepared by
50cm
coherent Bremsstrahlung from a diamond crystal. The
LH 2Target
asymmetry was again determined by detecting the decay
photons of the pion with the TAPS detector. Figure 3
0.4
0.3
0.2
504 BaF 2 Detector Modules 0.1
Fig. 2. One BaF module and the complete setup of the photon
0.0
spectrometer TAPS at the MAMI A2-Collaboration.
-0.1
(LET) [4], which derived from general principles a value ChPT
-0.2 DR
for the s-wave amplitude E0+ at threshold.
t to the data
These rst experiments showed a serious discrepancy -0.3
from the prediction of the low-energy theorem. The de- 0 20 40 60 80 100 120 140 160 180
velopment of the formalism of Heavy-Baryon Perturba- cms
0 / deg
tion Theory resolved this puzzle by showing, that the
LET value is only the leading term in a slow converging Fig. 3. The polarized photon asymmetry (from [3]).
H. Merkel: Experimental tests of Chiral Perturbation Theory 131
+
MAMI 1.31 0.08 10.02 0.2 10.5 0.2 13.1 0.1
ReE0+ / 10 /m
SAL 1.32 0.05 10.26 0.1
-3
ChPT 1.16 10.33 0.6 11.0 0.6 11.7 0.6 -0.6
DR 1.22 10.54 11.4 10.2
-0.8
0
-1.0
+
+
-1.2 n
Ethr =151.4 MeV
-1.4
145 150 155 160 165
p n p E / MeV
Fig. 4. Rescattering graph. Fig. 5. Real part of the s-wave multipole E0+ (see [3] for
details).
In the threshold region, the cross section can be further A = |S0+ |2 + |P5 |2 ,
decomposed into s- and p-wave multipoles: B = 2 Re (S0+ P4 ) ,
T () = p/k A + B cos + C cos2 , C = |P4 |2 |P5 |2 ,
L () = p/k A + B cos + C cos2 , G = Im E0+ P5 + S0+ P2 ,
T L () = p/k (D sin + E sin cos ) , H = Im (P1 P5 + P4 P2 ) .
T T () = p/k F sin2 ,
# 3.1 Unpolarized experiments
Q2
T L () = p/k (G sin + H sin cos ) ,
2 A rst electroproduction experiment at NIKHEF [10] es-
timated the s-wave cross section at threshold, a second
where p/k is the phase space ratio of pion CM momentum
experiment [11] tried to extract p-wave amplitudes in ad-
and photon CM equivalent momentum, and the angular
dition by measuring the pion emission angle in plane left
coe cients are combinations of two s-wave and ve p-wave
and right from the virtual photon direction. While these
multipoles:
pioneering works were in good agreement with ChPT, the
1 predictive power was not full exploited, since the data has
A = |E0+ |2 + |P2 |2 + |P3 |2 ,
2 to be analyzed with assumptions from theory.
B = 2 Re (E0+ P1 ) , A more complete experiment was performed at
1 MAMI [12] at the three-spectrometer setup of the A1
C = |P1 |2 |P2 |2 + |P3 |2 , collaboration. In electro production, the recoil proton is
2
detected in coincidence with the scattered electron. Very
D = Re (E0+ P5 + L0+ P2 ) , close to threshold, the boost by the Lorentz transforma-
E = Re (P1 P5 + P4 P2 ) , tion from the center-of-mass system to the laboratory sys-
1 tem focuses the full solid angle of the center of mass system
F = |P2 |2 |P3 |2 , into a narrow cone.
2
H. Merkel: Experimental tests of Chiral Perturbation Theory 133
1.0 2.0
E0+ / 10 /m
1.5
tot [b/sr]
0.5 1.0
2 0.5
0.0 0.0
0.00 0.05 0.10 0.00 0.05 0.10
3.0 4.0
W = 2.5 MeV W = 3.5 MeV
0 2.5
0 5 10 3.0
tot [b/sr]
W / MeV 2.0
0
L0+ / 10 /m
1.5 2.0
-3
1.0
1.0
0.5
0.0 0.0
0.00 0.05 0.10 0.00 0.05 0.10
-1
Q2 [GeV2/c2] Q2 [GeV2/c2]
Fig. 8. The total cross section [14] versus photon virtuality
(only MAMI data). The solid line shows the calculation in
HBChPT [13].
-2
0 5 10 Figure 8 summarizes the result of this experiment in
W / MeV comparison with the real photon data and the data at
Fig. 7. The extracted s-wave multipoles of the MAMI data [12] Q2 = 0.1 GeV2 /c2 . While the photon point and the higher
(solid squares) in comparison with the NIKHEF data [11] (open Q2 point was included in the t of HBChPT, the middle
circles) and a calculation in HBChPT [13] (line). Q2 point was not. As can be seen, there was a clear dis-
crepancy which has to be claried.
2.5 0.00
ChPT
Maid
2.0 DMT
0.10
LT [ub/sr]
0 [ub/sr]
1.5
0.20
1.0
0.30 ChPT
0.5 Maid
DMT
0.0 0.40
0 10 20 30 40 0 10 20 30 40
W [MeV] W [MeV]
0.40 0.5
ChPT
0.30 Maid
DMT 0.0
LT [ub/sr]
0.20
TT [ub/sr]
0.10 0.5
0.00
1.0 ChPT
0.10 Maid
DMT
0.20 1.5
0 10 20 30 40 0 10 20 30 40
W [MeV] W [MeV]
Fig. 9. Separated structure functions [15]. The solid line is the calculation in HBChPT [13], the dashed line the Maid model [16],
and the dash-dotted line the DMT model [17].
stated before, this calculation overestimates the dieren- tional to the imaginary part of L0+ multiplied by the large
tial cross section. Even more striking, the derivation of the (1232) multipole M1+ . Since it is very small, one has to
transverse-transverse interference is considerable, which is dene the helicity asymmetry
proportional to the dierence of the p-wave multipoles P2
and P3 , which are reproduced in good quality in photo- (h = 1) (h = 1)
AT L ( = 90 , = 90 ) =
production. (h = 1) + (h = 1)
For comparison, the phenomenological model
2(1 )T L ()
MAID [16] is included (dashed line). For the struc- =
ture functions, this model is able to describe the data sets T () + L () T T ()
roughly. This model is basically a global t to the existing
data sets in photo- and electro-production and can be to reduce the systematic errors. The expected energy
seen as check of the consistency of the data with the other structure is given by the unitary cusp at the opening of
existing data sets, however data at this photon virtuality the + threshold.
and energy are scarth. The dynamical Dubna-Mainz- Figure 10 shows this asymmetry. Again, the calcula-
Taipeh model (DMT) [17] is shown as dashed-dotted line tions in HBChPT and the models MAID and DMT are
and is in good agreement with the data. included in the graph. This small asymmetry enhances
the dierences between the models, and as can be seen
clearly, only the DMT model is able to describe the data.
3.3 Helicity asymmetry
By using polarized electrons, the fth structure function 4 Coherent production from the deuteron
LT can be extracted. This structure function has the
multipole contents T L (90 ) Im[L0+ (3 E1+ M1+ + The 0 production on the neutron is in principle a strong
M1 ) E0+ (2 L1+ L1 )], i.e. it is basically propor- prediction of HBChPT, since by the 0 production from
H. Merkel: Experimental tests of Chiral Perturbation Theory 135
nel with proton and neutron in the nal state. The con-
tribution of this channel was calculated in a model and
subtracted from the data. The dierential cross section
was extrapolated to the threshold to extract the thresh- 0
old s-wave amplitude. 100 110 120 130 140 150 160
2
A threshold value of Ed = (1.45 0.09)103 /m was Missing Mass [MeV/c ]
extracted. This value agreed within the error bar with a
Fig. 11. Missing mass resolution for the reconstruction of the
calculation in HBChPT [20], which gave a prediction for neutral pion [21].
the coherent s-wave amplitude only.
136 The European Physical Journal A
d/d [nb/sr]
d/d [nb/sr]
fit 1
15 fit 2 15 30
10 10 20
5 5 10
0 0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
cos() cos() cos()
d/d [nb/sr]
d/d [nb/sr]
30 30 30
20 20 20
10 10 10
0 0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
cos() cos() cos()
d/d [nb/sr]
d/d [nb/sr]
45 45 45
30 30 30
15 15 15
0 0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
cos() cos() cos()
d/d [nb/sr]
d/d [nb/sr]
80
60 60 60
40 40 40
20 20 20
0 0 0
-1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1
cos() cos() cos()
Fig. 12. The MAMI data on coherent electro production from deuteron [21] for three dierent values of the photon polarization
in comparison with the calculations in HBChPT [23]. The upper curve shows the t with xed Ld , the lower curve shows the
t with two free low-energy constants. The dashed lines show the previous results from ref. [24].
section with the usual assumptions of s- and p-waves with were extracted to
the known energy dependence was performed to extract
the threshold s-wave amplitudes. It has to be stressed, |Ed | 0.42 103 /m ,
that this t ignores completely the deuteron structure and |Ld | = (0.50 0.11) 103 /m .
nal state interaction, leading to a large systematic errors
in the extracted amplitudes. Since Ld is very large, for Ed The transverse multipole Ed was within agreement
only an upper limit could be extracted. The amplitudes with the calculations as at the photon point, while the
H. Merkel: Experimental tests of Chiral Perturbation Theory 137
calculations overestimated the longitudinal multipole by The success of ChPT in the SU (2) sector encourages
a factor of 2. one to extend this program to the strangeness threshold.
The situation improved a lot by a new calculation by At MAMI a new kaon spectrometer is under construction,
the same group [23], which now used a complete multipole and kaon threshold production will be among the rst
decomposition and extracted s- and p-waves. By this, the experiments with this spectrometer.
comparison has not to be done on the level of the ques-
tionable extracted s-wave amplitudes, but can be done
directly with the dierential cross section. References
Figure 12 shows the result of their calculation in com- 1. E. Mazzucato et al., Phys. Rev. Lett. 57, 3144 (1986).
parison with the MAMI data. The plot shows two kind of 2. R. Beck et al., Phys. Rev. Lett. 65, 1841 (1990).
ts: the rst was performed by xing the Ld multipole to 3. A. Schmidt et al., Phys. Rev. Lett. 87, 232501 (2001).
the extracted value of the data set, reducing the number 4. P. de Baenst, Nucl. Phys. B 24, 633 (1970); I.A. Vain-
of parameters to a single low-energy constant. This t, shtein, V.I. Zakharov, Nucl. Phys. B 36, 589 (1972).
however, does not describe the data set completely. The 5. V. Bernard, N. Kaiser, J. Gasser, U.-G. Mei ner, Phys.
second t was performed with two free parameters and Lett. B 268, 291 (1991); V. Bernard, N. Kaiser, U.-G.
describes the data. Mei ner, Z. Phys. C 70, 483 (1996).
6. J.C. Bergstrom et al., Phys. Rev. C 53, R1052 (1996).
7. M. Fuchs et al., Phys. Lett. B 368, 20 (1996); A.M. Bern-
stein et al., Phys. Rev. C 55, 1509 (1997).
8. V. Bernard et al., Eur. Phys. J. A 11, 209 (2001).
5 Summary and outlook 9. O. Hanstein et al., Phys. Lett. B 399, 13 (1997).
10. T.P. Welch et al., Phys. Rev. Lett. 69, 2761 (1992).
While the data basis on pion threshold production is now 11. H.B. van den Brink et al., Phys. Rev. Lett. 74, 3561 (1995).
quite large, there are still a number of unsolved prob- 12. M.O. Distler et al., Phys. Rev. Lett. 80, 2294 (1998).
lems. The strong Q2 -dependence of the total cross sec- 13. V. Bernard, N. Kaiser, U.-G. Mei ner, Nucl. Phys. A 607,
tion (g. 8) seems to be unlikely and might indicate the 379 (1996); 633, 695 (1998)(E).
large systematic errors of the data sets as quoted by the 14. H. Merkel et al., Phys. Rev. Lett. 88, 012301 (2002).
authors. 15. M. Weis, Doctorate Thesis, Mainz, 2003.
To resolve this puzzle, a dedicated experiment was per- 16. D. Drechsel et al., Nucl. Phys. A 645, 145 (1999); S.S.
formed by MAMI, where three dierent Q2 values were Kamalov et al., Phys. Lett. B 522, 27 (2001).
measured within one experiment and with emphasis on 17. S.S. Kamalov et al., Phys. Rev. Lett. 83, 4494 (1999);
reducing the systematic error [25]. This experiment is un- Phys. Rev. C 64, 032201 (2001).
18. M. Rekalo, E. Tomasi-Gustafsson, Phys. Rev. C 66, 015203
der analysis, but preliminary data seems to indicate, that
(2002).
the Q2 = 0.1 GeV2 /c2 data set is somewhat high. Since
19. J.C. Bergstrom et al., Phys. Rev. C 57, 3203 (1998).
that data set is included in the t of HBChPT, a ret of 20. S.R. Beane et al., Nucl. Phys. A 618, 381 (1997).
the theory seems to be necessary as soon as the new data 21. I. Ewald et al., Phys. Lett. B 499, 238-244 (2001).
set is published. 22. V. Bernard, H. Krebs, U.-G. Mei ner, Phys. Rev. C 61,
All cited calculations are based on the heavy-baryon 58201 (2000).
formalism. Meanwhile, better regularization schemes are 23. H. Krebs, V. Bernard, U.-G. Mei ner, Eur. Phys. J. A 22,
on the market and an improvement of the calculations can 503-514 (2004).
be expected. Especially in electroproduction, where the p- 24. H. Krebs, V. Bernard, U.-G. Mei ner, Nucl. Phys. A 713,
waves are not yet calculated to the same order as in photo 405 (2003).
production, a considerable improvement is possible. 25. J. Garc a Llongo, Diploma Thesis, Mainz, in preparation.
Eur. Phys. J. A 28, s01, 139 148 (2006)
DOI: 10.1140/epja/i2006-09-015-4 EPJ A direct
electronic only
/
Published online: 16 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. In 1953, it was decided to build a 500 MeV electron synchrotron in Bonn. It came into op-
eration 1958, being the rst alternating gradient synchrotron in Europe. After ve years of performing
photoproduction experiments at this accelerator, a larger 2.5 GeV electron synchrotron was built and set
into operation in 1967. Both synchrotrons were running for particle physics experiments, until from 1982
to 1987 a third accelerator, the electron stretcher ring ELSA, was constructed and set up in a separate
ring tunnel below the physics institute. ELSA came into operation in 1987, using the pulsed 2.5 GeV
synchrotron as pre-accelerator. ELSA serves either as storage ring producing synchrotron radiation, or as
post-accelerator and pulse stretcher. Applying a slow extraction close to a third integer resonance, external
electron beams with energies up to 3.5 GeV and high duty factors are delivered to hadron physics experi-
ments. Various photo- and electroproduction experiments, utilising the experimental set-ups PHOENICS,
ELAN, SAPHIR, GDH and Crystal Barrel have been carried out. During the late 90s, a pulsed GaAs
source of polarised electrons was constructed and set up at the accelerator. ELSA was upgraded in order to
accelerate polarised electrons, compensating for depolarising resonances by applying the methods of fast
tune jumping and harmonic closed orbit correction. With the experimental investigation of the GDH sum
rule, the rst experiment requiring a polarised beam and a polarised target was successfully performed at
the accelerator. In the near future, the stretcher ring will be further upgraded to increase polarisation and
current of the external electron beams. In addition, the aspects of an increase of the maximum energy to
5 GeV using superconducting resonators will be investigated.
PACS. 29.20.-c Cyclic accelerators and storage rings 29.20.Lq Synchrotrons 29.27.Hj Polarised beams
was constructed from dierent tubes glued together with Table 1. Main parameters of the 500 MeV synchrotron.
AralditTM . A van de Graa accelerator generated a 3 MeV
electron beam, which was injected into the synchrotron. Focusing type AG
The transfer beamline was calculated and set up by a PhD Number of basic periods 9
student [5], and a 60 kV electrostatic deector served for Basic lattice 1/2D-F -1/2D
Maximum eld at design orbit 1T
particle injection. Its housing was nally completely built
Maximum eld gradient 10 T/m
out of AralditTM in order to overcome the beam deteri-
Repetition frequency 50 Hz
orating eects of eddy currents, generated by the fringe Circumference 16.45 m
elds of the bending magnets. Orbit radius 1.70 m
Figure 1 shows a photograph of the 500 MeV syn- Revolution frequency 18.12 MHz
chrotron and part of its injection beamline. The main pa- No. of betatron oscillations per turn 2.4
rameters of the machine are given in table 1 (see also [6]). Gap height at orbit 6 cm
After overcoming a couple of additional throwbacks, Momentum compaction factor 0.16
the accelerator came into operation in 1958, being the Coils, no. of turns per magnet unit 98
rst strong focusing synchrotron operational in Europe. AC voltage / current 11670 V / 215 A
This happened to some extent unexpectedly, and actually DC voltage / current 100 V / 180 A
the institute was not prepared to perform scientic ex- Accelerating frequency 163.1 MHz
periments at the accelerator at this time. So Paul sent Peak voltage per turn 2.5 kV
a member of his machine group to the USA in order Number of accelerating cavities 6
to learn about physics experiments at the weak focusing Injection energy 3 MeV
1 GeV synchrotron of the California Institute of Technol- Injector type van de Graa
ogy. In the following years, a number of pilot experiments, Vacuum chamber ceramics
mainly on the photoproduction of pions o protons and Pressure (4 oil diusion pumps) 10 6 Torr
deuterons, were carried out. For the rst time the recoil
neutron polarisation was measured. Two rotating targets
produced external photon beams by the process of brems- based on the construction of the machine and scientic
strahlung. The energy of the external photon beam was experiments at this accelerator.
measured by means of tagging counters using one bend-
ing magnet of the synchrotron as dispersive element to
determine the momentum of the scattered electrons. 3 The 2.5 GeV synchrotron
The 500 MeV synchrotron was operated until 1984.
The total operation time exceeded 100000 hours. More In 1963, it became clear to the physicists perform-
than 150 diploma and doctoral theses were carried out ing experiments at the 500 MeV synchrotron that the
W. Hillert: The Bonn Electron Stretcher Accelerator ELSA: Past and future 141
Fig. 2. The Bonn 2.5 GeV electron synchrotron, close to its nal set-up.
experimental facilities of the Bonn accelerator laboratory Table 2. Main parameters of the 2.5 GeV synchrotron.
had to be extended in order to keep up with the develop-
ment of high-energy physics worldwide. A new electron ac- Focusing type AG
celerator with higher energy was considered, which should Number of basic periods 12
again be a university machine where scientic assistants Basic lattice O/2-F D-O/2
Field at 2.3 GeV 1.003 T
and students should participate in design and construc-
Field index n 22.26, 23.26
tion. At this time, the DESY synchrotron came close to
Repetition frequency 50 Hz
operation, the Cambridge accelerator was already operat- Circumference 69.6 m
ing and it was decided to build the NINA synchrotron. Orbit radius 7.65 m
In order to contribute to lling the gap between these Revolution frequency 4.3074 MHz
6 GeV machines and the existing ones around 1 GeV, the No. of betatron oscillations per turn 3.4
Bonn group decided to build an electron synchrotron with Gap height at orbit 6 cm
a maximum energy of 2.5 GeV. Maximum of momentum compaction 1.647 m
At the end of 1963, the state government agreed to Minimum of momentum compaction 0.81 m
support the new accelerator and the design of the ma- Coils, no. of turns per magnet unit 36
chine started. The general dimensions of the synchrotron Peak current at 1 T 1360 A
were xed by the limited area of approx. 30 60 m2 avail- Vacuum chamber ceramics
able between the two physics institute buildings. Therefore Accelerating frequency 499.67 MHz
a simple magnetic structure with twelve combined func- Voltage per turn 700 kV
RF peak power 40 kW
tion bending magnets based on a O/2-F D-O/2 lattice was
Number of accelerating cavities 2
chosen. In order to correct for chromatic eects, an addi-
Injection energy 25 MeV
tional sextupole component was introduced in the focusing
Injector type linac
and defocusing sectors of the bending magnets. To reach a Accelerating frequency (injector) 2998 MHz
short construction time the pole prole and the cross sec-
tion of these AG magnets were calculated on a computer,
using dierent two-dimensional relaxation methods which
had been developed shortly before [7,8]. In order to avoid dure at other labs, no magnet model was checked in ad-
corrections of quadrupole and higher order elds at the vance. The eects of the transition zones between the F-
time of injection, it was intended to inject at a eld of at and D-sectors and the various identical blocks of each sec-
least 0.01 T. From the beginning on a slow extraction of tor, which could not be modelled by two-dimensional com-
the electron beam was planned. puter simulations, were measured after the rst magnet
In 1964, the design was completed and the major parts had been assembled, and were compensated by well de-
were ordered. In 1965, the accelerator building and the signed exponentially shaped end blocks forming the mag-
laboratories were constructed. Manufacturing of the mag- net ends.
nets blocks was started based only on the performed com- At the end of 1966, almost all components were in place
puter simulations. In contrast to the usual design proce- and in March 1967, the rst electrons could be injected
142 The European Physical Journal A
and, without problems, accelerated up to 2.3 GeV. Accel- rst operation period from 1967 to 1984, the synchrotron
eration was performed by means of two cylindrical RF served as accelerator for particle physics experiments and
cavity resonators, which had been developed and con- as source for synchrotron radiation. A broad spectrum
structed in a special electroforming process by DESY, of physics experiments has been performed at this ma-
Hamburg [9]. A conventional television transmitter, built chine, starting from photo- and electroproduction of pseu-
by Telefunken, with a modied power amplier was used doscalar mesons o protons and deuterons, continuing
as RF power source. A linear accelerator, manufactured by with photoproduction of associated strangeness in K/ 0
Varian (Paolo Alto, California), served as injector, deliv- and nally including the measurement of the recoil nu-
ering a pulsed 25 MeV beam with a current of 250 mA at cleon polarisation. Since 1970, polarised solid state targets
an energy spread of 0.5 %. In July 1967, the accelerated (starting with polarised protons and later continuing with
electrons could be extracted with an e ciency of about polarised neutrons and deuterons) became available and
60 %. Slow beam extraction was performed by exciting a were also employed.
half integer horizontal betatron resonance with the help A oor plan of the experimental hall is shown in
of the nonlinear magnetic eld of a current strip. g. 3, presenting the experimental set-ups at the two syn-
The main parameters of the 2.5 GeV synchrotron are chrotrons.
listed in table 2 (see also [10]). Figure 2 shows a pho- The total operation time of the 2.5 GeV synchrotron
tograph of the machine, close to its nal set-up. In its amounted to 85000 hours for its rst operation period. In
W. Hillert: The Bonn Electron Stretcher Accelerator ELSA: Past and future 143
PHOENICS
SAPHIR
Magnetstromver-
Synchrotron
M Tagger
sorgung
Dipole (horizontal)
T
Dipole (vertical)
extraction septa
hadron Quadrupole
Sextupole
physics ELAN
superconducting DORIS cavity
Transformatoren,
Solenoid
Radio Frequency
NHV1
10 kV
Trafo
stretcher ring
0.5 - 3.5 GeV bending magnet
booster beamlines for
SR experiments
synchrotron BN3
BN2
0.5 - 1.6 GeV injection septa
DESY cavity
BN1
EKS
half cell of stretcher ring BN0
LINAC 1
(20 MeV) pol. e - M Q BPM
source SR
(120 keV) beamline
Mott polarimeter
electron
gun FZK laboratory
electron detector
gun
tests
LINAC 2
(26 MeV)
0m 5m 10 m 15 m
Fig. 4. Floor plan of the ELSA accelerator laboratory, indicating the rst experimental set-up of the time period 1989 1997.
total 210 diploma and doctoral theses were carried out and for excitation of a third integer betatron resonance
during this time. needed for slow beam extraction. Two long straight sec-
tions with vanishing dispersion are equipped with the ac-
celerating cavities and beam injection and extraction el-
4 The 3.5 GeV ELSA stretcher ring ements. Depending on the maximum beam energy cho-
sen, a single-cell resonator of DORIS type or two ve-cell
After more than ten years of experimental particle physics resonators of PETRA type are driven by klystron-based
at the 2.5 GeV synchrotron, it turned out that the qual- transmitters, operating at a frequency of 500 MHz and de-
ity of the experimental data was severely limited by the livering a maximum power of 40 kW and 250 kW, respec-
low duty factor of the pulsed synchrotron, which amounts tively. The vacuum system is based on thin wall (0.3 mm)
to about 5 % and could not be increased signicantly with oval tubes of stainless steel, whose rigidity is provided by
this type of accelerator. Following the suggestions of other 1 mm thick reinforcing ribs being brazed on the tubes.
laboratories around the world, it was decided to build a
A oor plan of the ELSA facility is shown in g. 4, rep-
pulse stretcher ring, using the 2.5 GeV synchrotron as in-
resenting the rst experimental set-up (1989 1997). The
jector [11]. A proposal was made in 1979, which was re-
main parameters of the stretcher ring are listed in table 3.
vised in 1981 and nally accepted at the end of 1981. After
several years of planning and designing, the construction ELSA can be operated in three dierent modes: the
of the new ELectron Stretcher Accelerator ELSA started stretcher mode, the post accelerator mode and the storage
in 1982. mode.
To avoid additional costs, the 500 MeV synchrotron Stretcher mode: Single pulses from the booster syn-
was dismantled in 1984 in order to use the existing ac- chrotron are injected into the stretcher ring at a maximum
celerator building for experiments at the stretcher ring. rate of 50 Hz. Using a slow extraction at a third integer
The new accelerator was placed in a separate tunnel sys- tune, an external electron beam of constant intensity is
tem constructed below the physics institute. ELSA is a obtained for the time between two injections. The max-
separated function machine of simple FODO-type, which imum energy is limited to 1.6 GeV by the beam transfer
provides radiation damping of the horizontal betatron os- from the synchrotron to the stretcher ring.
cillations necessary for beam storage and allows a wide Post accelerator mode: After injection of several pulses,
range variation of the betatron tune. In addition to the the accumulated electron beam is accelerated to the re-
dipoles and quadrupoles, a total number of twelve sex- quired energy and then extracted slowly. The maximum
tupoles were installed for correction of chromatic eects energy obtainable is 3.5 GeV, limited by the dipole magnet
144 The European Physical Journal A
Number of sextupoles 12 Expert programs Expert programs More than 12000 parameters
Gap height (dipole magnets) 5 cm Distributed
data base
Distributed
data base in data base, 19 permanently
Momentum compaction factor 6.3 % running expert programs.
Natural emittance (3.5 GeV) 0.9 mm mrad TCP/IP
UDP, RPC
Natural energy width (3.5 GeV) 0.09 %
Current (dipoles) at max. energy 3015 A Local Local Local Equipment control layer
data base data base data base
Vacuum chamber stainless steel 32 VME computers with
real time OS VxWorks (WindRiver)
Accelerating frequency 499.67 MHz Driver Driver Driver 5 PCs with Linux
Energy loss per turn (3.5 GeV) 1.22 MeV
TCP/IP, HDLC,
Maximum RF power 250 kW BUEP64, RK512
Accelerating cavities transmitter 1 1 of type DORIS Device interface layer
Accelerating cavities transmitter 2 2 of type PETRA 12 PLCs
57 MACS computers
(monoboard computers
developed in Bonn)
IEC bus,
power supply and the RF generated acceleration voltage. analog and digital I/O...
The macroscopic duty factor depends on the ramping Power Power measuring
supplies supplies instruments Equipment
speed and the at top time and scales inversely propor- etc. etc. etc.
graphical user interface combines machine steering, diag- intensity and polarisation were enlarged using circularly
nostics and data analysis in one integrated environment. polarised UV laser light for photoionisation of unpolarised
In combination with the control system, a new accelerator atoms. Based on this method (Fano eect), two sources
timing system, based on programmable delay units, was of polarised electrons were built in Bonn; one operating
also set up. The improved systems were the key to precise with caesium, the other with rubidium gas [26,27]. At the
and exible tracking of the main magnets on the energy Bonn 2.5 GeV synchrotron, with the help of the rubidium
ramp and enabled a fast ramping operation (maximum source, polarised electrons were accelerated for the rst
speed about 7 GeV/s) of ELSA. In addition they allow for time worldwide in a synchrotron [28].
joining of several injections from the synchrotron in ELSA The work on sources of polarised electrons was con-
with adjustable overlap, thus making a nearly perfect ho- tinued after the construction of ELSA, using the photo-
mogeneous lling of the stretcher ring possible. Since the emission of GaAs crystals pumped with circularly po-
successful implementation of the post accelerator mode larised laserlight. After more than ve years of work set-
at the end of 1994, ELSA was no longer operated in the ting up a suitable vacuum system and a pulsed titanium
stretcher mode. sapphire laser of su cient power, another source, based on
Dierent experimental set-ups were supplied with an a GaAs-AlGaAs superlattice photocathode was brought
external electron beam. into operation in 1997 [29]. This 120 keV source was used
The PHOENICS experiment was performed until the to investigate the eects of depolarising resonances in the
end of 1996, utilising a polarised frozen spin target and stretcher ring [30]. During these studies, it turned out
carrying out pion and eta photoproduction o protons and that reliability, life time and beam transfer e ciency of
deuterons (see [20] and references therein). this source would be insu cient for hadron physics exper-
The ELAN experiment (1988 1997) performed elec- iments.
troproduction and electrodisintegration experiments, de- The situation improved with the construction and as-
tecting the scattered electrons with a magnetic spectro- sembling of a new source, adapted for the operation with a
meter [21]. second linear accelerator, which had been moved from the
It was nished in August 1997 and followed up by an university of Mainz to Bonn and installed at the 2.5 GeV
experiment requiring a beam of circularly polarised pho- synchrotron. This 50 keV source [31] is based on an invert-
tons and a polarised proton target in order to investigate ed-geometry electron gun, operated in space charge limi-
the contributions to the GDH sum rule in the energy range tation in order to suppress the spiking of the free running
accessible at ELSA (see [22] and references therein). The ashlamp-pumped titanium sapphire laser. Special care
GDH experiment was carried out in the years 2000 2002 was taken in construction and set-up of the gun and the
after ELSA had been successfully upgraded for the oper- transfer beamline to reach and maintain a low base pres-
ation with polarised electrons (see sect. 5). sure (1011 mbar) and extremely low partial pressures of
The SAPHIR spectrometer was operated from 1991 to poisoning gas species (1014 mbar) by application of dier-
1999, carrying out photoproduction experiments and con- ential pumping [32]. To improve the gun vacuum and con-
centrating on the detection of charged particles. A large sequently the lifetime of the photocathodes, heat cleaning
variety of reactions has been studied with this detector, and activation of the photocathodes are carried out in a
reaching from photoproduction of associated strangeness load-lock system, which in addition allows to change crys-
in K/ 0 to the production of vector mesons , and tals without breaking the vacuum of the gun (see g. 6).
(see [23] and references therein).
The 50 keV source was operated from 2000 to 2003
SAPHIR was disassembled in 1999 and followed up
with a Be-InGaAs/AlGaAs strained layer superlattice
by a new experimental set-up (CB@ELSA), based on the
photocathode [33] for machine studies and the GDH ex-
4 photon detector Crystal Barrel which had been moved
periment, emitting a peak current of 100 mA in rectangu-
in 1997 from LEAR/CERN to Bonn and equipped with
lar 1 s long electron pulses with about 80 % beam polari-
an inner scintillating ber detector for charged particle
sation and demonstrating photocathode lifetimes of more
detection and triggering. During the rst data-taking pe-
than 2000 hours.
riod of the CB@ELSA experiment from 2001 to 2003, pro-
ton, deuteron and solid state targets were used in photo- To prevent depolarisation during acceleration in the
production experiments (see [24] and references therein). circular accelerators due to spin precession around the
A linearly polarised photon beam, produced by coherent guiding eld of the dipole magnets, the electron spins,
bremsstrahlung o a goniometer-aligned diamond crystal, originally orientated longitudinally at the source, have to
was set up in 2002 for this experiment and has been rou- be rotated in order to point perpendicularly to the ac-
tinely available since then. celerator plane. This rotation is performed by a 90 de-
gree electrostatic bend in the low energy beamline from
the source to the linac. An additional Larmor precession
5 Polarised electrons in the linear accelerator, caused by its focusing solenoid
lenses, is compensated by additional double solenoids in
The production of polarised electron beams has been stud- the injection beamline, which allow to vary the spin rota-
ied in Bonn for more than 35 years. Already in 1969, a tion angle while their focusing strengths remain xed. Af-
source of polarised electrons, based on the photoionisa- ter beam extraction out of the stretcher ring, a supercon-
tion of polarised lithium atoms, was set up [25]. Beam ducting solenoid rotates the spin back into the accelerator
146 The European Physical Journal A
Operation Parameters
Acceleration voltage 48 kV
Repetition rate 50 Hz
Pulse length 1 s
Pulse current 100 mA
Beam polarisation 80 %
Laser spot size (diameter) 8 mm
-11
Pressure in gun chamber 10 mbar
Beam lifetime > 2000 h
Fig. 6. Set-up of the inverted source of polarised electrons and the load-lock system.
L
0.8
Longitudinal
Horizontal
PTagger / PELSA
superconducting Vertical
solenoid 0.6 synchrotron
bending
magnets
side-band
0.4 Px
J
quadrupole
Px, Py, Pz
magnets
0.2
T 0
1 1.5 2 2.5 3 3.5
Energy / GeV
Horizontal orbit
10
ga+0 5
Q
x / mm
Qsp = ga 0
x -5
Qz 0 20 40 60 80 100 120 140 160
DQz s/m
Vertical orbit
t0 3
2
1
Dt1 Dt2 t
z / mm
0
-1
-2
Q
-3
0 20 40 60 80 100 120 140 160
5 s/m
10-
8-g
12-
4
6-g
ga-2
2
ga+
ga+
ga+
ga
ga
Qz orbit.
at
Injection
4
0 1 2 3 4 5 6 7 8 ga Source
Source
0 440 880 1750 2640 3500 E / MeV
0 150 400 640 t / ms
Fig. 9. Fast tune jumping of all intrinsic resonances in ELSA.
Energy / GeV
but it is required to correct for all resonances strong
enough to cause a signicant depolarisation. Fig. 11. Achieved beam polarisation in the stretcher ring.
Due to the fast ramping speed and high superperiodic-
ity, no signicant depolarisation is observed in the booster
synchrotron if the beam is transferred to ELSA at ener- variation and measuring the polarisation response of the
gies below the third imperfection resonance at 1.32 GeV. extracted beam with a Mller polarimeter.
A dierent situation shows up for the stretcher ring. It With all correction methods applied successfully, po-
turned out from numerical simulations that at least nine larised electrons can be accelerated up to 3.2 GeV. At en-
resonances in ELSA are strong enough to produce signi- ergies higher than 2.0 GeV some polarisation loss is ob-
cant depolarisation. Three techniques are applied to avoid served due to incomplete resonance compensation (see
depolarisation: g. 11).
In case of intrinsic resonances, the crossing speed is
enhanced with the help of two pulsed betatron tune jump
quadrupoles [37], thereby shifting the vertical betatron 6 Future plans
tune by Qz 0.1 in a t2 10 ms long triangular
pulse with a rise time of t1 = 4 s (see g. 9). Starting in 2006, a new experimental set-up of CB@ELSA
The strengths of the imperfection resonances are re- (see g. 12) will come into operation, utilising a po-
duced with a dynamic correction of the closed orbit during larised beam and a polarised nucleon frozen spin target
the energy ramp. Vertical and horizontal displacements of for photoproduction experiments in the framework of the
the beam in the quadrupoles are measured by 28 mon- SFB/TR 16 Subnuclear Structure of Matter , funded by
itor stations [38] and corrected with 19 horizontal and the Deutsche Forschungsgemeinschaft (DFG). With this
21 vertical corrector magnets [39]. The following correc- set-up stronger demands will be put on the quality of the
tion scheme is used: The beam is stored at the energies external electron beam. In order to enhance the beam po-
of the imperfection resonances and the optimal correc- larisation and the external current at maximum beam en-
tions are determined. Afterwards a linear interpolation ergy, a number of improvements are planned in the near
between these corrections is applied during the energy future, the most important of which are the following:
ramp. The remaining distortions are generally smaller The 32 old quadrupole vacuum chambers will be re-
than 0.2 mm rms (see g. 10). placed by new watercooled ones, equipped with improved
Further reduction of the resonance strengths is monitor stations (capacitive pickups) and clearing elec-
achieved by correction of specic harmonics of closed orbit trodes for ion clearing.
distortions relevant for a single imperfection resonance. The existing 40 corrector magnets will be removed and
This method is based on an empirical determination of 60 newly designed magnets will be installed, together with
two amplitude factors for each resonance by parameter four-quadrant power supplies developed at the institute.
148 The European Physical Journal A
/
Published online: 15 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. The Mainz Microtron MAMI is a cascade of three racetrack microtrons, delivering since 1991
a high-quality 855 MeV, 100 A cw-electron beam for nuclear, hadron and radiation physics experiments.
An energy upgrade of this facility to 1.5 GeV by adding a Harmonic Double-Sided Microtron (HDSM)
as a fourth stage is well underway and rst beam is expected during the rst half of 2006. A detailed
description of the multiple recirculation scheme with normal conducting accelerator structures, the basis
for the reliable operation of MAMI, is given and the historical development from MAMI A to MAMI B
is described. The natural advancement to MAMI C by realizing a polytron of the next higher order, the
HDSM, is covered in the last section and a rst glimpse into the future of MAMI is given.
PACS. 29.20.-c Cyclic accelerators and storage rings 41.75.Lx Other advanced accelerator concepts
41.85.Lc Beam focusing and bending magnets, wiggler magnets, and quadrupoles
Fig. 2. Example of a eld map of a RTM 180 -dipole magnet before and after applying the surface correction coils.
the number of turns, Eout/in is the output/input beam Originally, the second stage RTM2 was planned with
energy, respectively. an end energy of only 100 MeV with one klystron TH2075
Numerical BBU-calculations together with rf-measure- feeding RTM1 and RTM2 [7]. However, because of a strong
ments on the MAMI biperiodic accelerating structures [12] demand to surpass already with this machine distinctly
showed, that by staggered T M110 -detuning of the linac the pion production threshold, it was decided to add a
sections the threshold current can be shifted distinctly second klystron and accelerate in RTM2 by a factor of
above the maximum design current of 100A. thirteen from 14 MeV to 180 MeV. The cost increase for
the larger end magnets could be lowered by using the iron
of the Heidelberg Cyclotron in use from 1943 to 1973 [14].
3 From MAMI A to MAMI B This setup (Van-de-Graa + RTM1 + RTM2 MAMI A)
delivered from July 1983 to October 1987 about 70% of its
Following these design principles MAMI was realised as a 18700 hours of beam time for hadron and nuclear physics
3-RTM-cascade between 1979 and 1990. Since the skep- experiments. The maximum achieved beam parameters
ticism that a machine with many recirculations could be were 187 MeV beam energy and a current of 65 A. Its
build a proof of principle was requested. Therefore, a operation was funded since 1984 by a Collaborative Re-
14 MeV stage (MAMI A1) was built for testing and op- search Centre (SFB201, Medium energy physics with the
timising the rf-structure and rf-control, the B-eld cor- electromagnetic interaction ) and the main components
rection by the surface coil technique and, quite advanced were bought via the HBFG funding.
at that time, a complete computer control using steer-
ing algorithms. It was set into operation in March 1979 A somewhat weak point of MAMI A was the cheap
and later used as the rst RTM (RTM1) of MAMI A and Van-de-Graa injector. Its maximum usable voltage of
MAMI B. This machine was already used for rst physics only 2.1 MV ( = 0.981) caused a migration of the op-
experiments [13] from November 1979 on. Only the klys- erating phase in RTM1 from +15 to 22 , resulting in
tron TH2075 and the Van-de-Graa injector were bought, a reduced longitudinal acceptance and stability. This ad-
otherwise it consisted completely of in-house-made or used verse eect was enhanced by the high sensitivity of the sta-
components as, e.g., the end magnets from DESY, Ham- bility of the high voltage of the Van-de-Graa to any im-
burg. pact of background radiation. Because this background
152 The European Physical Journal A
Fig. 4. Annual operation time of MAMI, according to ma- Fig. 5. Simplied scheme of the measurement of the absolute
chine setup (tuning and development), polarised and unpo- beam energy (top) and beam energy uctuations (bottom) in
larised beam time, for the years 1991 to 2005. It should be RTM3.
noted that in 2001 a half year shutdown took place for the
preparation of the beam lines for MAMI C. Of the remaining
4428 hours the MAMI was operating for 4277 hours, i.e. 97%.
tem [16], allowing to computer control a wide diversity of
parameters and feedback loops. Among the most impor-
radiation increased with increasing beam current, the volt- tant monitors are the low-Q rf-cavities on the linac axes
age instabilities of the Van-de-Graa were the reason for of each RTM. They allow by injection of 10 ns diagnos-
the limit of max. 65 A beam current. Moreover, the tic beam pulses during machine setup, actually realised
mediocre vacuum conditions and bad accessibility of the as 10 ns blackouts in the cw-operation, to supervise the
high voltage terminal were evidently prohibitive for any transverse positions and the phase and intensity of the
operation of a GaAs-photocathode source of polarised beam for each recirculation individually at the entrance
electrons. and exit of the linac. This information make quick and
Therefore, when transferring MAMI A as the injector e cient correcting actions possible by a machine model
for RTM3, the Van-de-Graa was replaced by a 3.5 MeV implemented in the computer control system [17].
linac designed and built in-house [15] with high energy The beam prole is viewed turn by turn and in the
stability ( 1 keV; = 0.992). At this energy the phase transfer lines between the RTMs via a synchrotron radi-
migration in RTM1 is only 12 to 22 . ation camera system. Very helpful for fast tuning of the
The nal scheme realised as MAMI B, set up in new matching of the beam parameters is a synchrotron radia-
halls from 1987 to 1990 with rst operation in August tion camera with high magnication looking through the
1990, is shown in g. 12, and its main data are given in axes of the linacs of RTM2 and RTM3. All 51 (RTM2)
table 1. and 90 (RTM3) turns have to overlap and any mismatch
Apart from its excellent beam parameters (table 1), can easily be detected and globally corrected. A system of
the machine showed an extraordinarily stable and reliable many T M110 -rf-cavities allows a control of the beam po-
operation. The beam time over the years of operation since sition in diagnostic-pulse and also cw mode (above 1 A
1991, classied for machine setup and polarised and un- beam current) down to a few m.
polarised operation for experiments, is shown in the his- The RTM conguration easily permitted in RTM3 the
togram of g. 4. The high e ciency of MAMI is due on installation of two control setups extremely valuable for
one side to the inherent properties of the RTM, but also precision experiments (g. 5). The distance between the
to a considerable extent to a sophisticated monitor sys- return pipes is such that small 4 2.45 GHz = 9.8 GHz
A. Jankowiak: The Mainz Microtron MAMI Past and future 153
Table 1. Main parameters of MAMI. MAMI C comprises the injector linac, MAMI A, MAMI B and the HDSM being con-
structed.
Injector RTM1 RTM2 RTM3 HDSM
General
injection / extraction energy (total) [GeV] 0.511/3.97 10 3
3.97/14.86 10 3
14.86/180 10 3
0.180 / 0.855 0.855 / 1.5
number of turns 18 51 90 43
total power consumption [kW] 92 92 220 650 1400
Rf-System
energy resp. energy gain / turn [MeV] 3.5 0.599 3.24 7.50 16.58-13.66
frequency [GHz] 2.4495 2.4495 2.4495 2.4495 4.8990|2.4495
linac length (electrically) [m] 4.93 0.80 3.55 8.87 8.57|10.1
number of sections / klystrons 3/1 1/1 2/2 5/5 8/4|5/5
power dissipation / beam power [kW] 33.2 / 0.35 7.9 / 1.1 48.4/16.6 102.5 / 67.5 299 / 65
power consumption [kW] 90 90 180 450 1000 a
Magnet-System
ux density (within the gap) [T] 0.1026 0.5550 1.2842 1.53-0.95
gap height [cm] 6 7 10 8.5-13.9
min./max. de ection radius [m] 0.129-0.482 0.089-1.083 0.467-2.216 2.23-4.60
iron / copper weigth of the magnets [t] 4 / 0.2 90 / 2.3 900 / 11.6 1000 / 27.4
number of corrector magnets 40 72 204 360 2 172 + 2 6
number of quadrupoles and solenoids 20 2 4 4 24
power consumption [kW] 2 2 40 200 400
Beam-Parameters
energy spread (1) [keV] 1.2 1.2 2.8 13 110 b
norm. emittance hor. / vert. (1) [ 10 6 m] 0.05 / 0.04 0.07 / 0.07 0.25 / 0.13 13 / 0.84 27 b / 1.2 b
standard-energies for experiments 180MeV 195-855MeV 0.855-1.5GeV
in steps of 15MeV in steps of ca. 15MeV
a
Including the power consumption of one matching section between RTM3 and HDSM.
b
Simulation with SYTRACE, a particle tracking program including eects of stochastic emission of synchrotron radiation photons.
- T M010 - and T M110 -resonators can be inserted there The development of the polarised source [19] after
(T M110 -cavities at 720 and 855 MeV and T M010 -cavities MAMI ran with a clean linac as injector is listed in the
at 315, 420, 510, 570 and 855 MeV). So rstly, with the following time table:
T M110 -position monitor and the known distance between 1992 1995: First experiments for the A3-collabora-
this monitor and the linac axis, one can use one 180 - tion, beam current I = 5 A, polarisation P = 30
dipole with its NMR-controlled eld strength and pre- 40%. The 100 keV source was installed at ground
cisely measured eld map as a sensitive spectrometer with level meaning a 14 m long beam line to the injector
large 2.2 m bending radius to determine the absolute beam linac [20].
energy to 2 104 (140 keV at 855 MeV). Secondly, 1995: Introduction of strained layer cathodes, I =
with an additional 9.8 GHz T M010 -cavity on the extrac- 2 A, P = 75%.
tion path and monitoring its phase dierence to one of the 1997: The source moves to the accelerator hall, which
return-pipe T M010 -resonators, one can precisely measure allows a much easier and reproducible injection of the
changes of the length of the last half turn after extrac- beam into the linac. Installation of a 2f-prebuncher
tion of the respective energy. The corresponding energy with 4.9 GHz [21], which allows for 145 longitudinal
change of the electron bunches is given by eq. (2) and the phase space acceptance instead of the design value of
sensitivity reads 40 [15], I = 10 A, P > 75%.
rf /2 61.2 mm 1998: Introduction of the so called synchro-laser for
= = 8.16 mm/MeV, (6) pulsing the source, which allows for 90% transmis-
E/turn 7.5 MeV
sion of the precious polarised electrons, I = 20 A,
which corresponds to 96 phase per MeV at 9.8 GHz. With P = 80%.
a resolution of 0.1 at 9.8 GHz energy changes of about 2001: Introduction of the mask activation technique,
1 keV, corresponding to 1.2 106 at 855 MeV, are de- which strongly reduces losses of electrons starting at
tected. A further increase of resolution seems not to be the cathode due to stray light and thus improves the
reasonable at the moment since uctuations of the beam vacuum conditions at the cathode. The charge ex-
direction are producing signal levels of about the same tracted in one run was increased from 22 C to 115 C.
amount. By feeding back the energy signal to the linac From then on several weeks of continuous operation at
phase, it is possible to routinely provide this energy sta- high current I > 20 A were possible.
bility of 1 keV during physics experiments. Of course, 2003 2005: A Wien lter as spin-rotator at 100 keV
correct tuning and a su cient stability of the longitudinal electron energy directly behind the polarised source
Q-value are key to the well functioning of the system [18]. was installed. This allows for a much easier adjust-
This setup was of fundamental importance for the parity- ment of the beam polarisation at the target for all ex-
violating electron scattering experiments, where the cross perimental stations and all energies. Before, the spin
5
section change with Ebeam and must be measured with a direction at the experiments was controlled by tuning
relative precision of 106 . the MAMI end energy making use of the gyromagnetic
154 The European Physical Journal A
Fig. 6. Detailed scheme of the Harmonic Double Sided Microtron (HDSM) for MAMI C.
anomaly of the electron. Currents of I > 30 A with well proven and tested technology applied for the RTMs:
polarisations of P > 85% are now possible. normal conducting rf-accelerator structures and iron core
magnets with normal conducting excitation coils.
Apart from steady improvements of the degree of polarisa-
The latter point clearly implied that one could not
tion and the lifetime of the photocathodes, the installation
realise MAMI C as a fourth RTM. With iron core mag-
of the harmonic 2f-prebuncher at the injector linac and the
nets one cannot increase the eld strength very much
rf-synchronised laser for up to 90% transmission e ciency
beyond the 1.3 T of RTM3. The size and weight of the
made it possible to increase the MAMI operation with po-
two such 180 end magnets would grow with the cube
larised electrons from 20% to now 60% of the total beam
of the maximum energy, i.e. to formidable weight of
time satisfying the demands of the experiments.
450 t(1.5/0.855)3 = 2430 t each. However, the RTM with
one linac is not the only possible microtron. Already since
1979 H. Herminghaus and K.H. Kaiser developed ideas
4 The Harmonic Double-Sided Microtron and designs for higher order microtrons called Poly-
(HDSM) as the fourth stage of MAMI trons , as multi-turn recirculators with strong phase fo-
cusing ([22,23,24]).
In 1999 a new Collaborative Research Centre (SFB443, At the bicyclotron or Double-Sided Microtron
Many-body structure of strongly interacting systems ) (DSM) (see the scheme in g. 6) the pole face area is
was founded, which for its second stage physics program reduced by a factor of ( 2)/ compared to an RTM.
demanded an electron beam of 1.5 GeV. The ideas for up- Therefore, a 1500 MeV DSM has roughly the same mag-
grading the MAMI energy had to consider as boundary nets weight as an 850 MeV RTM. As the next step one
conditions that the excellent beam quality and reliabil- must consider, however, that the dynamic coherence con-
ity of MAMI B must be preserved, that the new fourth dition changes and for a DSM is given by
stage had to t into the existing buildings and that the
research with the existing MAMI B had to go on without ecB ecB
E/turn = n 2rf = n rf , (7)
any longer shutdown periods. Moreover, considering the 2( 2) 2
limited manpower capacity of the institute and the tight
time schedule envisaged, it was evident that one had to and naturally one will take n = 1 for the lowest possi-
base the new accelerator stage on the expertise of the insti- ble path lengthening of 2 rf /turn. With the parameters
tute. So very early the decision was taken to stay with the of RTM3 (B = 1.28 T, rf = 0.1224 m) one would need
A. Jankowiak: The Mainz Microtron MAMI Past and future 155
Fig. 12. Floor plan of MAMI C. The installation up to RTM3 (MAMI B) has not been changed essentially since 1990 (A2:
Tagger / A4: Parity Violation / X1: Radiation Physics, till 2000 in the HDSM Hall).
of the HDSM. The contract was given to THALES. Con- at the next prototype by Ti-coating the nose cones of
sidering both bidders would have caused an increase of the the klystron resonators, but unfortunately through the
total costs of this system of more than 50%. But obviously higher surface resistivity thermal problems occurred. Fi-
the problems for fabricating a power klystron at this high nally, with a total delay of 26 months all tubes needed for
frequency were underestimated. It took a long series of MAMI C were delivered, fortunately, with a production
prototypes partly damaged by trivial technical failures guarantee for the TH2166 tube till 2010 by THALES.
and 27 months compared to the anticipated 12 months However, with somewhat reduced specications (50 kW,
delivery time, before the desired tube was in house and 45% e ciency and multipactor freeness only for some-
could be used for the power tests of many other 4.9 GHz what restricted operating conditions), which is just safely
components (rf-structures, circulators, special waveguide adequate for the 1.5 GeV operation of the HDSM.
components). These tests were successfully performed in
All other high power components, e.g. the circula-
2003. However, the prototype TH2166-tube showed strong
tors by AFT and the water loads from Spinner, worked
multipacting discontinuities on its transfer curve and was
satisfactorily from the beginning. The two 30 kV / 27 A
therefore not qualied for precision operation at the ac-
klystron power supplies built by BRUKER, Wissenbourg,
celerator. THALES could solve the multipacting problem
were successfully operated in several longterm tests,
158 The European Physical Journal A
Fig. 13. Measured vertical B-eld of HDSM dipole No.2 normalised to the ideal eld gradient (top), and the construction
drawing of the corresponding correction coil (bottom). The dark quadratic area at the lower left of the eld map is due to a
piece of parallel pole faces necessary for an NMR-probe for precise eld regulation.
Table 2. Main parameters of the HDSM dipoles. was awarded to the French company USINOR2 . Aside the
eld strength [T] 1.53 - 0.95
promising manufacturing capabilities of this company the
gap distance [mm] 85 - 139
main argument was, that only USINOR oered to pro-
mech. length of front edge [m] 7
duce the magnets essentially of only two symmetric pieces
usable length of front edge [m] 6.5 (upper and lower piece), which is clearly the favourable
iron weight [t] 250 geometry to avoid any discontinuities perpendicular to the
coils copper prole outside [mm2 ] 12 12 pole edge. The magnet pieces, each weighing 125 t, were
coils cooling duct diameter [mm] 8 casted out of high permeable iron and then machined at
number of windings 2 256 the company SFAR, a subcontractor of USINOR. This
current/voltage [A/V] 212/340 machining procedure for a high quality and precise surface
copper weight [t] 6.85 of the partly concave pole pieces was worked out in close
collaboration with IKPh. Due to the complicated pole ge-
ometry it was expected, that for the nal eld correction
of the magnets to the 104 level not only symmetric, but
especially during the high power conditioning and test- also asymmetric eld errors (resulting in unwanted eld
ing of the fully installed and commissioned 2.45 GHz linac components in the plane of beam acceleration) must to
of the HDSM. be corrected. Based on the well proven concept of surface
Naturally, beside the task to design and build up the correction currents [35], a procedure had been developed
worldwide rst 4.9 GHz cw linac, the manufacturing of the which allows to extract the symmetric and asymmetric
four 90 -bending magnets with eld gradient presented eld components by a simultaneous measurement of the
the second highly critical challenge for the completion of vertical magnetic eld in and 25 mm out of the midplane
the HDSM. The mechanical and magnetic design of these of the magnets, and to construct surface correction coils
dipoles was completely done at IKPh. The main goal was which compensate both errors simultaneously ([36,37]).
to get magnets with excellent eld quality at minimum size The rst magnet was delivered end of 2001 and all four
(existing halls) and iron consumption ([33,34]). The main magnets were nally in place end of 2002. The contract for
parameters of these magnets are given in table 2. The
2
call for tender started in 1999 and in 2000 the contract Today SFAR STEEL (Le Creusot, France).
A. Jankowiak: The Mainz Microtron MAMI Past and future 159
the manufacturing of the excitation coils was awarded to and ready for operation, whereas the installation of the
the company SIGMAPHI (Vannes, France). They intro- 4.90 GHz linac, after the nal delivery of the 10 needed
duced a special bi-lar winding technology, which allowed accelerator sections, has just started. The next step is
to realise optimum heat distribution within the coils, to the installation of the two recirculation path vacuum sys-
avoid internal brazing and to choose reasonable power sup- tems and the completion of the injection and extraction
ply parameters [34]. Both guarantees a high reliability over beam lines. The rst operation of the HDSM is expected
the lifetime of the accelerator. Each magnet is fed by an in the rst half of 2006. After a period of commission-
individual, highly stabilised power supply (478 V, 260 A, ing in diagnostic pulse mode with low beam power (10 ns,
short term/long term stability: 3 ppm/10 ppm) manufac- high-intensity bunch trains with a repetition rate of max.
tured by DANFYSIK (Jyllinge, Denmark). By feeding 10 kHz), very soon the rst physics experiments will be
back the reading of NMR-probes to the PS, the eld of started since all upgrades of the beams lines, the photon
each magnet is stabilised to better than 105 . tagger and the spectrometers has been nished.
It took till September 2003 to nish all magnet eld
measurements. To explore the capabilities of the mag- Many people worked together to realise the very success-
nets for a later energy upgrade, these measurements were ful operation of MAMI over the last 25 years. Here I just
not only done at the nominal eld of 1.53 T but also at want to mention the certainly most important ones: Hel-
1.64 T = 1.61 GeV and even 1.71 T = 1.67 GeV. In g. 13 mut Herminghaus, who is the intellectual father of MAMI and
the measured eld of dipole No.2 (at the nominal eld of laid the strong foundations for this success story and Karl-
1.53 T normalised to the ideal eld gradient) is plotted. Heinz Kaiser, who overtook the responsibility for the MAMI
In the central area of the magnet the eld deviations are operation and development in the early 1990s and set the
already in the order of 104 , a clear proof of the excel- guidelines for the future of MAMI: the design and installation
lent work done by USINOR/SFAR. As a further result of the Harmonic Double-Sided Microtron. The construction
and operation of MAMI would not have been possible with-
of this high manufacturing precision, the analysis of the
out Hans Euteneuer who is, amongst others, responsible for
asymmetric eld errors of the magnets showed, that the
the developments in the rf-eld. His great help for the prepa-
transverse components are well below 1 mT. A rough es- ration of this manuscript must be stressed as my last point.
timation of the inuence of the resulting vertical beam
deections of 0.1 mrad to max. 0.35 mrad leads to an ac-
ceptable coupling of only a few percent between the hor-
izontal and vertical phase spaces. So it was decided to do References
the nal correction only for symmetric eld errors, result-
ing in much simpler identical upper and lower correction 1. J. Haimson, Linear Accelerators, edited by P.M. Lapos-
coils. In the lower part of g. 13 a sketch of one of this cor- tolle, A.L. Septier (Amsterdam, 1970) p. 415.
rection coils, manufactured by water jet cutting of a 3 mm 2. D. Husmann, IEEE Trans. Nucl. Sci. NS-30, No. 4, 3252
(1983).
thick aluminium plate, is shown. With these pairs of coils
3. L. Harwood, Proceedings of PAC2003, Portland, OR, USA
the desired eld accuracy of 2 104 was easily achieved
(2003) p. 586.
for all four magnets. One can clearly see, that most of the
4. E.M. Moroz, Sov. Phys. Dokl. 1, 326 (1956).
correction must be done near the corners of the magnet, 5. A. Roberts, Ann. Phys. (N.Y.) 4, 115 (1958).
because here quite large eld decays exist. This behaviour 6. B.H. Wiik et al., Linear Accelerators, edited by P.M. La-
was already predicted by TOSCA-simulations and are due postolle, A.L. Septier (Amsterdam, 1970) p. 553.
to the triangular cut necessary to t the magnets as far as 7. H. Herminghaus et al., Nucl. Instrum. Methods 138, 1
possible into the corners of the accelerator hall. It turned (1976).
out, that even at the design eld level of 1.53 T this eld 8. H. Herminghaus et al., Nucl. Instrum. Methods 187, 103
decay leads to deection errors of up to 2.2 mrad at low (1981).
electron energies. Because it reaches far into the fringe 9. P. Axel et al., IEEE Trans. Nucl. Sci. NS-24, No. 3, 1133
eld region, it cannot be corrected by surface correction (1977).
coils alone. Therefore, at the entrance and exit corner of 10. O. Hanson, Charlottesville Conference Paper Q (1979).
each dipole individually designed vertical iron shims at- 11. H. Herminghaus et al., Nucl. Instrum. Methods 163, 299
tached to its front face are necessary. Together with the (1979).
steering magnets on the return paths and the linac axis 12. H. Euteneuer et al., Proceedings of LINAC84, Seeheim,
they will provide a proper angle and position correction of Germany (1984) p. 394.
the beam [37]. Because with increasing eld of the dipoles 13. M. Begemann et al., Nucl. Instrum. Methods 201, 287
the eld decay at the magnet corners gets much stronger, (1982).
a later energy upgrade of the HDSM, based on the ex- 14. U. Schmidt-Rohr, Die Deutschen Teilchenbeschleuniger
periences gained during the operation at 1.5 GeV, will (U. Schmidt-Rohr, Heidelberg, 2001) p. 144.
most probably require the construction and installation 15. H. Euteneuer et al., Proceedings of EPAC88, Rome, Italy
of a fully new set of correction coils and iron shims. (1988) p. 550.
16. H. Euteneuer et al., Proceedings of LINAC92, Ottawa,
Presently, all four dipole magnets are aligned and Canada (1992) p. 356.
equipped with their individual set of correction coils and 17. H.J. Kreidel, PhD Thesis, KPH 12/87, University of
vacuum chambers. The 2.45 GHz linac is commissioned Mainz, Mainz, Germany (1987).
160 The European Physical Journal A
18. M. Seidl, Proceedings of EPAC2000, Vienna, Austria 28. H. Euteneuer et al., Proceedings of LINAC86, Stanford,
(2000) p. 1930. CA, USA (1986) p. 508.
19. K. Aulenbacher et al., Journal AIP, Vol. 675 (2002) 1088. 29. H. Euteneuer et al., to be published in Proceedings of
20. H. Steens PhD Thesis, KPH 01/94, University of Mainz, EPAC06, Edinburgh, GB (2006).
Mainz, Germany (1994). 30. H. Euteneuer et al., Proceedings of EPAC00, Vienna, Aus-
21. V.I. Shvedunov et al., Proceedings of EPAC96, Barcelona, tria (2000) p. 1954.
Spain (1996) p. 1556. 31. A. Jankowiak et al., Proceedings of LINAC04, Lubeck, Ger-
22. K.H. Kaiser, Proceedings of the Conference on Future many (2004) p. 842.
Possibilities for Electron Accelerators, Charlottesville, VA, 32. G. Faillon et al., Proceedings of LINAC86, Stanford, CA,
USA (1979) V-1. USA (1986) p. 122.
23. H. Herminghaus et al., Proceedings of LINAC81, Santa Fe, 33. U. Ludwig-Mertin et al., Proceedings of EPAC98, Stock-
N.M., USA (1981) p. 260. holm, Sweden (1998) p. 1931.
24. H. Herminghaus, Nucl. Instrum. Methods A 305, 1 (1991). 34. A. Thomas et al., Proceedings of EPAC02, Paris, France
25. S. Ratschow, PhD Thesis, KPH 02/00, University of (2002) p. 2379.
Mainz, Mainz, Germany (2000). 35. H. Herminghaus, Proceedings of EPAC88, Rome, Italy
26. A. Jankowiak et al., Proceedings of EPAC02, Paris, France (1988) p. 1151.
(2002) p. 1085. 36. M. Seidl et al., Phys. Rev. STAB 5, 062402 (2002).
27. J. Herrmann et al., Proceedings of PAC99, New York, USA 37. F. Hagenbuck et al., Proceedings of EPAC04, Lucerne,
(1999) p. 2915. Switzerland (2004) p. 1669.
Eur. Phys. J. A 28, s01, 161 171 (2006)
DOI: 10.1140/epja/i2006-09-017-2 EPJ A direct
electronic only
/
Published online: 23 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. An extended experimental program to investigate the Gerasimov-Drell-Hearn (GDH) sum rule
and related partial reaction cross sections on proton and neutron has been carried out by the GDH collab-
oration at the electron accelerators MAMI (Mainz) and ELSA (Bonn). The GDH sum rule connects the
helicity-dependent photoabsorption cross section with the anomalous magnetic moment of the nucleon. The
GDH collaboration has measured the total cross section of circularly polarised photons with longitudinally
polarised protons at MAMI and ELSA to check this sum rule experimentally. In addition partial reaction
channels like pion, double pion and eta production were determined. This provides new information on the
helicity-dependent excitation spectrum of the nucleon. With the help of partial wave analyses it is possible
to extract new, complementary information on the broad, overlapping resonances in this energy region.
The double polarisation observable E measured in this experiment enhances the smaller multipoles via
interference terms. The analysis of our data provides a new possibility to study the photon couplings to
the nucleon resonances, especially above the resonance where many properties of the observed states
(e.g., coupling constants, branching ratios, helicity amplitudes) are only poorly known. In this paper we
present several new results from our measurements on polarised proton and deuteron targets.
PACS. 16.60.Le Meson production 14.20 Baryon resonances with S = 0 25.20 Photoproduction
Fig. 1. The Mainz Microtron MAMI. The GDH experiment was performed in the A2 tagger hall.
polarized -strip
target
polarized
-beam
shower
trigger plates
1m
Fig. 2. Overview of the GDH experimental setup at MAMI. The polarised photon beam is coming along the axis of the 3 He/4 He
refrigerator of the polarised frozen spin target.
new and up to now inaccessible information on partial- mum energy of 855 MeV. In our rst successful data taking
wave amplitudes. Besides the measurements with the po- period in 1998 (g. 3) we typically had a degree of polari-
larised proton target, we have also performed an extended sation of about 75%, in the 2nd period in the year 2003 for
series of experiments with polarised deuteron targets in or- the neutron runs this has been improved to about 80%. A
der to extract information on the neutron and thus on the dedicated experimental apparatus (g. 2) including a po-
isospin dependence of the helicity structure. larised solid target and a detector with full angular accep-
tance was installed in the A2 Hall. The photons were pro-
duced by bremsstrahlung in the A2-Glasgow-Mainz tag-
2 Experimental setup ging facility (g. 4), which rstly determines the photon
The MAMI accelerator (g. 1) with its source of polarised energy by the help of 352 scintillation counters with a res-
electrons, based on the photoeect on a strained GaAs olution of approximately 2MeV at 855MeV primary beam
crystal, routinely delivers polarised beams with a maxi- energy and secondly measures the degree of polarisation
A. Thomas: The Gerasimov-Drell-Hearn sum rule at MAMI 163
3 Results
3.1 Results on the GDH-Integral and the forward spin
Fig. 8. The time dependence of the target polarisation during polarizibility integral on the proton
the rst data-taking period in 1998.
Figure 10 shows our data for the helicity dierence of
the total cross section measured at MAMI in the rst
The relaxation time for the proton spins was about phase of the GDH experiment [10]. The data are compared
200 hours. Consequently we could take data for typically to predictions from the partial wave analyses SAID and
2 days before repolarizing the target, for example three MAID. The negative values of the cross section dierence
cycles of polarising the target and data taking during the (3/2 1/2 ) in the threshold region are due to the dom-
relaxation of the target can be seen in g. 8. The repo- inance of the E0+ multipole in the single + production
larizing time for the target (approximately 3 hours) was channel. The excitation of the -resonance (P33 (1232))
used to rell the 400l liquid-helium buer containers for prefers the 3/2 cross section at energies around 300 MeV.
the cryostat and the polarising magnet. The movement It is excited by a strong M1+ transition with only 2.5%
of the polarising solenoid and the detector took around E1+ admixture. At higher energies the double pion pro-
15 minutes. duction plays an important role and is not represented in
The cylindrical detector DAPHNE (Detecteur a the theoretical curves.
grande Acceptance pour la Physique photoNucleaire Ex- In g. 11 the experimental running GDH integral
perimentale) [9] was especially designed for handling multi (right-hand side of eq. (1)) is displayed and compared
particle nal states by provision of a large solid angle (94% to the model predictions. The integration starts at
A. Thomas: The Gerasimov-Drell-Hearn sum rule at MAMI 165
800 250
3/2 - 1/2 (barn)
100
200
50
0
0
-200
-50
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
E (MeV) E (MeV)
Fig. 10. The total cross section dierence (3/2 1/2 ) on 1 H Fig. 11. The running GDH integral obtained in this work
is compared to previous results [11] (open circles) and to the starting at 200 MeV is compared to the model predictions.
predictions of the HDT [12], SAID [13] and UIM [14] analyses. Only statistical errors are shown.
Only statistical errors are shown.
unpolarized
800MeV [21]
the single + production up to a photon energy of Since the unpolarised cross section of eq. (3) is given by
800MeV [22] the sum of absolute squares of multipoles, only a few dom-
the double pion production channels 0 + [23] and inant partial waves can be condently evaluated. On the
0 0 [24] other hand, the sensitivity to some the weaker multipoles
helicity dependent production [25]. is greatly enhanced by measuring the polarised cross sec-
tion 31 of eq. (4). In particular new interference terms
Besides the importance of this data to check the GDH
appear, e.g., between E1+ and M1+ , which are directly
sum rule and to measure the forward spin polarizibility as
related to the (1232) or P33 (1232) excitation in the low-
shown in the previous section, new information about the
energy region or between E2 and M2 , which are directly
nucleons excitation spectrum can be extracted. This fea-
related to the D13 (1520) excitation.
ture can be understood from the multipole decomposition
of the N N total cross section. In the following we
use the pion multipole notation, where E and M denote 3.3 The (1232)-resonance region
the electric or magnetic character of the incoming pho-
ton and the indices l describe the coupling of the pion There has been an extended program in dierent labo-
angular momentum l and the nucleon spin to the total an- ratories to increase our knowledge about the EMR ratio
gular momentum J = l 1/2. In the (1232)-resonance for the (1232)-resonance [27,28]. The determination of
region the measured helicity asymmetry is sensitive to the the double-polarisation observable E provides new, com-
E1+
ratio of the multipoles EM R = M 1+
. In the second res- plementary information to clarify this question. In gs. 15
onance region (500 MeV 900 MeV), where several and 16 we have compared our data with predictions from
overlapping states are present, e.g., P11 (1440), D13 (1520), the multipole analysis MAID2000 using EM R values of
S11 (1535), the helicity dependent observables are partic- 2.5%, 0% and +2.5%. The data are well reproduced with
ularly sensitive to the behaviour of the electromagnetic an EM R = 2.5% [20].
A. Thomas: The Gerasimov-Drell-Hearn sum rule at MAMI 167
Ds[mb]
200
-2.5%
150 gp np+ 0%
+2.5%
GDH98
100
50
-50
-100
-150
200 300 400 500
Eg[MeV]
Fig. 15. Sensitivity of the GDH observable 3/2 1/2 for the
EMR ratio of the (1232) resonance for single + production.
The curves were produced using MAID2000.
Ds[mb]
350 gp pp0 -2.5%
0%
300 +2.5%
GDH98
250
200
150
100
50
3.4 The second resonance region Additional channels are needed to disentangle the reso-
nances at higher nucleon excitation energies. Double pion
Since DAPHNE also has a moderate e ciency for neu- photoproduction is particularly important for the study
tral particle detection we can distinguish in our data all of the second resonance region, where the P11 (1440) state
the contributing partial reaction channels in the 2nd reso- with its unclear origin is located, since almost 50 % of the
nance region. Figure 19 shows our results for the single 0 total photoabsorption cross section can be attributed to
production. The agreement of the MAID2000 partial-wave the N channels. In the MAMI B energy range up to
analysis with our data could be improved signicantly by 800MeV we are presently analysing our data for the he-
changing the parametrisation of the multipoles E2 and licity dierence 3/2 1/2 of the double-pion production
M2 that drive the D13 -excitation by approximately 20%. channel + . The results for the + 0 and 0 0 chan-
In addition g. 20 shows our preliminary results [29, nels have already been published [23,24]. Comparing our
22] for single + production at higher energies. Although results with predictions from dierent theoretical mod-
the + channel is mainly produced via intermediate els [30,31] gives a new insight into the double pion pro-
D13 (1520) excitation, a cusp structure can be observed duction mechanism, specially the role of the D13 (1520),
close to the production threshold in the 1/2 channel. the P11 (1440) and the (1700) are under discussion.
168 The European Physical Journal A
(d/d)3/2-(d/d)1/2 [b/sr]
5
2.5
0
-2.5
560 MeV 580 MeV 600 MeV
-5
4
2
0
-2
620 MeV 640 MeV 660 MeV
-4
4
2
0
-2
680 MeV 700 MeV 720 MeV
-4
4
2
0
-2
740 MeV 760 MeV 780 MeV
-4
0 100 0 100 0 100
*
[deg]
3.6 photoproduction
Fig. 27. The new dilution refrigerator for the Crystal Ball
detector.
acknowledge the excellent support of the MAMI and ELSA 17. L. Tiator, Proceedings of GDH2002 (World Scientic, Sin-
accelerator groups. This work was supported by the Deutsche gapore, 2003).
Forschungsgemeinschaft (SFB 201, SFB 443, Schwerpunkt- 18. N. Bianchi, E. Thomas, Phys. Lett. B 450, 439 (1999).
programm 1034, and GRK683), the INFN-Italy, the FWO 19. S. Simula et al., Phys. Rev. D 65, 034017 (2002); private
Vlaanderen-Belgium, the IWT-Belgium, the UK Engineering communication.
and Physical Science Council, the DAAD, JSPS Research Fel- 20. J. Ahrens et al., Eur. Phys. J. A 21, 323 (2004).
lowship, and the Grant-in-Aid (Specially Promoted Research) 21. J. Ahrens et al., Phys. Rev. Lett. 88, 232002 (2002).
in Monbusho, Japan. 22. J. Ahrens et al., submitted to Phys. Lett. C.
23. J. Ahrens et al., Phys. Lett. B 551, 49 (2003).
24. J. Ahrens et al., Phys. Lett. B 624, 173 (2005).
References 25. J. Ahrens et al., Eur. Phys. J. A 17, 241 (2003).
26. D. Drechsel, Prog. Part. Nucl. Phys. 34, 181 (1995).
27. R. Beck, H.P. Krahn et al., Phys. Rev. Lett. 78 (1997).
1. S.B. Gerasimov, Yad. Fiz. 2, 598 (1965) (Sov. J. Nucl. 28. G.S. Blanpied et al., Phys. Rev. Lett. 79, 4337 (1997).
Phys. 2, 430 (1966)). 29. T. Rostomyan, PhD Thesis (Gent) (2005).
2. S.D. Drell, A.C. Hearn, Phys. Rev. Lett. 16, 908 (1966). 30. J. Nacher, Proceedings of NSTAR2001 (World Scientic
3. I. Karliner, Phys. Rev. D 7, 2717 (1973). Pub. Co, 2001) p. 189.
4. A. Sandor, Phys. Rev. D 50, R6681 (1994). 31. M. Vanderhaeghen, H. Holvoet, private communication,
5. Anselmino et al., Sov. J. Nucl. Phys. 49, 553 (1989). Mainz 2001.
6. H. Olsen, L.C. Maximon, Phys. Rev. 114, 887 (1959). 32. H. Arenhovel, The GDH for the deuteron, in Proceedings of
7. C. Bradtke et al., Nucl. Instrum. Methods A 436, 430 GDH2000, Mainz, edited by D. Drechsel, L. Tiator (World
(1999). Scientic, Singapore, 2001) p. 67.
8. H. Dutz et al., Nucl. Instrum. Methods A 340, 272 (1994). 33. A. Fix, private communication.
9. G. Audit et al., Nucl. Instrum. Methods A 301, 473 (1991). 34. M. Schwamb, private communication.
10. J. Ahrens et al., Phys. Rev. Lett. 87, 022003 (2001). 35. M. Martinez, PhD Thesis, University of Mainz (in prepa-
11. J. Ahrens et al., Phys. Rev. Lett. 84, 5950 (2000). ration).
12. O. Hansein et al., Nucl. Phys. A 632, 561 (1998). 36. St. Goertz et al., Nucl. Instrum. Methods A 526, 43
13. R.A. Arndt et al., Phys. Rev. C 66, 055213 (2002). (2004).
14. D. Drechsel et al., Nucl. Phys. A 645, 145 (1999). 37. O. Jahn, PhD Thesis, University of Mainz (2005).
15. H. Dutz et al., Phys. Rev. Lett. 91, 192001 (2003). 38. S. McGee, PhD Thesis, Duke University (in preparation).
16. H. Dutz et al., Phys. Rev. Lett. 93, 032003 (2004). 39. R. Novotny, IEEE Trans. Nucl. Sci. 38, 379 (1991).
Eur. Phys. J. A 28, s01, 173 183 (2006)
DOI: 10.1140/epja/i2006-09-018-1 EPJ A direct
electronic only
/
Published online: 26 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. A very successful experimental program with real photons has been achieved in 20 years of
operation at the Mainz Microtron (MAMI) facility. The dierent detector setups, like DAPHNE, TAPS
and the Crystal Ball are centered around the tagged photon facility the so-called Glasgow Tagger. From
the rich spectrum of results only a few highlights will be discussed here, the proton polarizabilities, the pion
polarizabilities, pion photoproduction close to the pion threshold and in the (1232)-resonance region.
PACS. 13.40.-f Electromagnetic processes and properties 13.60.Le Meson production 13.60.Fz Elastic
and Compton scattering 14.20.Dh Protons and neutrons
30
/(nb/sr)
25
20
lab
d /d
15
10
0
40
lab=107 lab=133
35
30
/(nb/sr)
25
20
lab
d /d
15
10
0
0 40 80 120 160 200
40
E /MeV
lab=155
35
30
/(nb/sr)
V. Olmos et al.
sum rule constraint and the value as follows from the
d /d
15
10 experiment by Zieger et al. [12]. The thick solid line shows the
5 result of the global t, eq. (4).
0
0 40 80 120 160 200
E /MeV
rule constraint leads to the following result:
Fig. 1. Measusred dierential cross-sections in the lab sys- = 12.1 0.3stat. 0.4syst. 0.3mod. ,
tem [9] compared with a dispersion relation calculation (solid (4)
line) [11]. = 1.6 0.4stat. 0.4syst. 0.4mod. ,
where the rst error denotes the statistical, the second the
The dierential cross sections obtained are plotted in systematic and the third the model-dependent one. The
g. 1. The systematic errors of 3% arise from uncer- results are summarized in g. 2 (contourplot) where the
tainties in the photon ux (2%) and the target density contours in the ( ) plane for 2min + 1 are plotted. In
(2%) combined in quadrature. The eective solid addition, the Baldin sum and the value obtained from the
angles were determined with Monte Carlo simulations. experiment by Zieger [12] are included.
Errors from uncertainties in the experiment geometry are
estimated to be 5%. With the help of the dispersion
relation approach the electromagnetic polarizabilities 3 Pion polarizability
of the proton can be extracted from the experimental
cross-sections. The procedure used in the analysis was The pion polarizabilities characterize the dynamical defor-
to take and as free parameters, and sometimes as mation of the pion in the electromagnetic eld. The values
well the constraint given by the Baldin sum rule. Using of the electric and magnetic pion polarizabilities de-
standard 2 minimization, the result obtained, when pend on the rigidity as a composite particle and provide
tting the MAMI/TAPS data alone without the sum rule important information of internal structure. Very dierent
constraint, is values for the pion polarizabilities have been calculated in
the past. All predictions agree, however, that the sum of
= 11.9 0.5(stat.) 1.3(syst.), (2) the two polarizabilities of the meson is very small.
= 1.2 0.7(stat.) 0.3(syst.). (3) On the other hand, the values of the dierence of the po-
larizabilities are very sensitive to theoretical models. For
The Baldin sum rule obtained from this result, example, investigations within the framework of the chiral
+ = 13.1 0.9 is in agreement with the value perturbation theory (ChP T ) predict ( ) 5.4 [13]
determined by the total photon absorption cross section. in one-loop calculations and 4.4 1.0 for two-loops [14]
A t to the existing low-energy Compton scattering (all values of the polarizabilities are given in units of
data including the new MAMI/TAPS data and the sum 104 fm3 ). The calculations in the extended Nambu-Jona
R. Beck: Experiments with photons at MAMI 175
A
B A
MWPC + FSD
-beam
LH2
C
B
TOF
1m MWPC+FSD
LH2
Fig. 3. Floor plan of the experimental setup showing the loca- Fig. 4. Enlarged view showing the details of the TAPS con-
tion of the detectors. A, B, C are TAPS blocks, MWPC+FSD guration.
show multi-wire proportional chambers and the forward scin-
tillation detector, TOF indicates the block of the neutron de-
tector bars, and LH2 stands for the liquid-hydrogen target in
The experiment discussed here has been performed at
its vacuum scattering chamber.
the continuous-wave electron accelerator MAMI B [25,
26] using Glasgow-Edinburgh-Mainz tagger photon facil-
Lasinio model with linear realization of chiral U (3) U (3) ity [1,2]. The quasi-monochromatic photon beam covered
symmetry [15] result in = = 3.0 0.6. The the energy range from 537 to 819 MeV with an intensity
application of dispersion sum rules (DSR) at xed value 6 105 /s in the tagger channel for the lowest pho-
of the Mandelstam variable u = 2 for calculation of this ton energy and average energy resolution of 2 MeV. The
parameter [16,17] leads to ( ) = 10.3 1.9. DSR at tagged photons entered a scattering chamber, containing
nite energy [8] gave the similar result: ( ) = 10.6. a 3 cm diameter and 11.4 cm long liquid hydrogen target
A calculation in the linear model with quarks and vector with Capton windows. The emitted photon , + me-
mesons included to one loop order predicted ( ) = son, and the neutron were detected in coincidence. The
20 [18]. An evaluation in the Dubna quark connement experimental setup is shown in g. 3.
model [19] results in ( ) = 7.05. The photons were detected by the spectrometer
Because there is no stable pion target, experimental TAPS [4], assembled in a special conguration (g. 4). The
information about the pion polarizabilities is not easy to TAPS spectrometer consists of 528 BaF2 crystals. Each
obtain. One has to investigate reaction channels, like scat- hexagonally shaped crystal is 250 mm long corresponding
tering high energy pion in the Coulomb eld of a heavy to 12 radiation lengths. All crystals were arranged into
nuclei or the radiative pion photoproduction. three big blocks. Two blocks (A, B) consisted of 192 crys-
The scattering of high energy pions o the Coulomb tals arranged in 11 columns and the third block (C) had
eld of heavy nuclei [20] has resulted in = = 144 crystals arranged in 11 columns. These three blocks
6.8 1.4 1.2. This value agrees with prediction of the were located in the horizontal plane around the target
dispersion sum rules but is about 2.5 times larger than the at angles 68 , 124 , 180 with respect to the beam axis.
ChP T result. The experiment of the Lebedev Institute on Their distances to the target center were 55 cm, 50 cm and
radiative pion photoproduction from the proton [21] has 55 cm, respectively. All BaF2 modules were equipped with
given + = 20 12. This value has large error bars and 5 mm thick plastic veto detectors for the identication of
shows the largest discrepancy with regard to the ChP T charged particles.
predictions. The attempts to determine the polarizability The neutrons were detected by a wide aperture time-
from the reaction suer greatly from theoreti- of-ight spectrometer (TOF) [27]. It consisted of 111 scin-
cal [22] and experimental [23] uncertainties. The most re- tillation detector bars of 50 200 3000 mm3 and 16
cent analysis of MARK II and Crystal Ball data [24] nds counters (10 230 3000 mm3 ) which were used as veto
no evidence for a violation of the ChP T predictions. How- detectors. The bars are made from NE110 plastic scin-
ever, even changes of polarizabilities by 100% and more tilator and each bar is read out on both ends by two 3
are still compatible with the present error bars. phototubes XP2312B. All bars were assembled in 8 planes
176 The European Physical Journal A
(nb)
d/ds1dt (nb/ )
1.2
4
14
1 12
0.8 10
0.6 8
0.4 6
4
0.2
2
0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
2 0 550 600 650 700 750 800
t/
E (MeV)
Fig. 5. The dierential cross section of the process p + n
averaged over the full photon beam energy interval and over s1 Fig. 6. The cross section of the process p + n integrated
from 1.5m2 to 5m2 . The solid and dashed lines are the predic- over s1 and t in the region where the contribution of the pion
tions of model-1 and model-2, respectively, for ( )+ = 0. polarizability is biggest and the dierence between the predic-
The dotted line is a t to the experimental data (see text). tions of the theoretical models under consideration does not
exceed 3%. The dashed and dashed-dotted lines are predic-
tions of model-1 and the solid and dotted lines of model-2 for
of a special conguration with 16 detectors in each, fol- ( )+ = 0 and 14 10 4 fm3 , respectively.
lowing one after another (g. 3). Such a neutron detector
allows to detect the neutrons in the energy region 10 mentum transfer. As seen from this gure, the theoretical
100 MeV with e ciency 30 50% and to determine their curves are very close to the experimental data. This means
energy with a resolution 10% using the neutron time that the dependence of the dierential cross section on the
of ight and the angle of the neutron emission measured square of the four-momentum transfer t which is basically
with a precision 2 3%. the kinetic energy of the neutron is well reproduced by us-
To detect the + meson two two-coordinate multi- ing the mentioned GEANT simulations for the e ciency.
wire proportional chambers (MWPC) and a forward scin- In a second step, the kinematic region where the po-
tillator detector (FSD), for getting a fast trigger signal, larizability contribution is maximal was investigated. This
have been developed and constructed. The MWPCs over- is the region 5m2 s1 < 15m2 and 12m2 < t < 2m2 .
lap angles in the laboratory system were = 2 20 , In the considered region of the phase space, the cross sec-
= 0
360
and were located under 0
with respect to tions of the process p + n integrated over s1 and t
the beam direction. are calculated according to model-1 and model-2 for two
The cross section of the process p + n has been dierent values of (). The obtained experimental cross
calculated in the framework of two dierent models. In sections and their theoretical predictions for ()+ = 0
the rst model (model-1) the contribution of all the pion and 14 104 fm3 are presented in g. 6. The error bars
and nucleon pole diagrams is taken into account using are the quadratic sum of statistical and systematic errors.
pseudoscalar pion-nucleon coupling [28]. For each model, we obtain
In the second model (model-2), the nucleon and the
pion pole diagrams without the anomalous magnetic mo- ( )+ = (12.2 1.6stat 3.3syst ) 104 fm3
ments of the nucleons, and in addition the contributions (model 1), (5)
of the resonances (1232), P11 (1440), D13 (1520), and ( )+ = (11.1 1.4stat 2.8syst ) 104 fm3
S11 (1535) are included.
(model 2). (6)
To control the model dependence of the result the kine-
matic regions were limited to regions where the dierence Averaging over the results of the two models, the nal
between model-1 and model-2 does not exceed 3% when result is obtained [29]:
( )+ is constrained to zero. First, a kinematic region
where the contribution of the pion polarizability is negligi- ( )+ = (11.6 1.5stat 3.0syst 0.5mod ) 104 fm3 .
ble, i.e. the region 1.5m2 s1 < 5m2 was analysed, where (7)
s1 is the squared pion-photon center-of-mass energy.
In g. 5, the experimental data for the dierential cross 4 Pion photoproduction in the threshold
section, averaged over the full photon beam energy in- region
terval from 537 MeV up to 817 MeV and over s1 in the
indicated interval, are compared to predictions of model- The photoproduction of pions near threshold has been a
1 (dashed curve) and model-2 (solid curve). The dotted topic of considerable experimental and theoretical activi-
curve is the t of the experimental data in the region of ties over the past years, ever since the results of the ex-
10m2 < t < 2m2 , where t is the squared pion mo- periments, performed in Saclay [30], Mainz ([31,32]) and
R. Beck: Experiments with photons at MAMI 177
/ b
2.5
The discrepancy could be explained by a calculation in
the framework of heavy-baryon chiral perturbation theory 2.0
(ChPT) [36], which showed that additional contributions 1.5
due to pion loops in 2 have to be added to the old LET. 1.0
Rened calculations within heavy-baryon ChPT [37] led
0.5
to descriptions of the four relevant amplitudes at threshold
by well-dened expansions up to order p4 in the S-wave 0.0
144 147 150 153 156 159 162 165 168
amplitude E0+ and p3 in the P -wave combinations P1 , E / MeV
P2 and P3 , where p denotes any small momentum or pion
mass, the expansion parameters in heavy-baryon ChPT. Fig. 7. Total cross sections for 0 photoproduction close to
To that order, three low-energy constants (LEC) due to threshold with statistical errors (without systematic error of
the renormalization counter terms appear, two in the ex- 5%) as function of incident photon energy (solid squares, this
pansion of E0+ and an additional LEC bP for P3 , which work ref. [38], open circles, ref. [33], open diamonds ref. [32]).
have to be tted to the data or estimated by resonance
saturation.
However, two combinations of the P -wave amplitudes, where d and d are the dierential cross sections for
P1 and P2 , are free of low-energy constants. Their expan- photon polarizations perpendicular and parallel to the re-
sions in converge rather well leading to new LETs for action plane dened by the pion and proton. The asym-
these combinations. Therefore, the P -wave LETs oer a metry is proportional to the dierence of the squares of
signicant test of heavy-baryon ChPT. However, for this P3 and P2 :
test the S-wave amplitude E0+ and the three P -wave com-
binations P1 , P2 and P3 have to be separated. This separa- q d()
() = (P32 P22 ) sin2 ()/ . (10)
tion can be achieved by measuring the photon asymmetry 2k d
using linearly polarized photons, in addition to the mea-
surement of the total and dierential cross sections. A measurement of the reaction p( , 0 )p [39] was per-
The dierential cross sections can be expressed in formed at the Mainz Microtron MAMI [40] using the
terms of the S- and P -wave multipoles, assuming that Glasgow/Mainz tagged photon facility [1,2] and the pho-
close to threshold neutral pions are only produced with ton spectrometer TAPS [4]. The MAMI accelerator de-
angular momenta l of zero and one. Due to parity and livered a continuous wave beam of 405 MeV electrons.
angular momentum conservation only the S-wave ampli- Linearly polarized photons were produced via coherent
tude E0+ (l = 0) and the P -wave amplitudes M1+ , bremsstrahlung in a 100 m thick diamond radiator [41,
M1 and E1+ (l = 1) can contribute and it is conve- 42] with degrees of polarization of up to 50%. The neu-
nient to write the dierential cross section and the pho- tral pion decay photons were detected in TAPS [43], an
ton asymmetry in terms of the three P -wave combinations array of 504 BaF2 detectors, which was built up around a
P1 = 3E1+ + M1+ M1 , P2 = 3E1+ M1+ + M1 and liquid-hydrogen target.
P3 = 2M1+ + M1 . The c.m. dierential cross section is The total and dierential cross sections were measured
over the energy range from 0 threshold to 168 MeV. Fig-
d() q ure 7 shows the results for the total cross section in com-
= (A + B cos() + C cos2 ()), (8) parison to ref. [33] and [32]. The results for the photon
d k
asymmetry are shown in g. 8 in comparison to the values
where is the c.m. polar angle of the pion with re- of ChPT [37] and to a prediction of a dispersion theo-
spect to the beam direction and q and k denote the retical calculation (DR) by Hanstein, Drechsel and Tia-
c.m. momenta of pion and photon, respectively. The co- tor [44]. The photon asymmetry was determined from all
e cients A = |E0+ |2 + |P23 |2 , B = 2 Re(E0+ P1 ) and the data between threshold and 166 MeV for which the
C = |P1 |2 |P23 |2 are functions of the multipole ampli- mean energy was 159.5 MeV. The theoretical predictions
tudes with P232
= 12 (P22 + P32 ). Earlier measurements of are shown for the same energy.
the total and dierential cross sections already allowed The values for the real and imaginary part of E0+
determination of E0+ , P1 and the combination P23 . and the three P -wave combinations were extracted via
In order to obtain E0+ and all three P -waves sepa- two multipole ts to the cross sections and the photon
rately and to test the new LETs of ChPT, it is necessary asymmetry simultaneously. The two multipole ts dier
to measure, in addition to the cross sections, the photon in the energy dependence of the real parts of the P -wave
asymmetry , combinations. For the rst t the usual assumption of a
d d behaviour proportional to the product of q and k was
= , (9)
d + d adopted (qk-t, 2 /dof = 1.28). The assumption made
178 The European Physical Journal A
Table 1. Results of both ts (qk-t and q-t) for Re E0+ at the 0 - and + -threshold (unit: 10 3 /m+ ), for the parameter
of Im E0+ (unit: 10 3 /m2+ ) and for the three combinations of the P -wave amplitudes (unit: q 10 3 /m2+ ) with statistical and
systematic errors in comparison to the predictions of ChPT [37, 45] (O(p3 )) and of a dispersion theoretical approach (DR, [44]).
This work ChPT DRa
qk-ta q-t
0
p
E0+ (Ethr ) 1.23 0.08 0.03 1.33 0.08 0.03 1.16 1.22
n +
E0+ (Ethr ) 0.45 0.07 0.02 0.45 0.06 0.02 0.43 0.56
2.43 0.28 1.0 5.2 0.2 1.0 2.78 3.6
P1 9.46 0.05 0.28 9.47 0.08 0.29 9.14 0.5 9.55
P2 9.5 0.09 0.28 9.46 0.1 0.29 9.7 0.5 10.37
P3 11.32 0.11 0.34 11.48 0.06 0.35 10.36 9.27
P23 10.45 0.07 10.52 0.06 11.07 9.84
a
Values of the P -wave combinations converted into the unit q 103 /m2+ .
0.4 0.0
this work: q k-t ChPT [8]
0.3 -0.2 this work: q-t DR [17]
0.2 -0.4
+
ReE0+ / 10 /m
-3
0.1 -0.6
0.0 -0.8
-0.1 -1.0
ChPT [8] +
-0.2 -1.2 n
DR [17] Ethr =151.4 MeV
this work t to the data
-0.3 -1.4
0 20 40 60 80 100 120 140 160 180 145 150 155 160 165
cms
0 / deg E / MeV
Fig. 8. Photon asymmetry for 0 photoproduction at 159.5 Fig. 9. Results for Re E0+ with statistical errors as a function
MeV photon energy with statistical errors (without systematic of incident photon energy E for an assumed energy depen-
error of 3%) as a function of the polar angle (solid line: t dence of the P -wave amplitudes proportional to q k (solid
to the data) in comparison to ChPT [37] (dotted line) and squares) and q (open squares) in comparison to ChPT [37]
DR [44] (dashed line). (dotted line) and DR [44] (dashed line).
0.06
MAMI / TAPS (E =298 MeV) 3/2 3/2
35.0 REM=(Im E1+ / Im M1+ ):
MAMI / DAPHNE 0.04 0 +
Multipolanalysis (p -, n -data)
30.0 LEGS (E =280 MeV)
/ ( b/sr)
0.02 Hanstein(DR)
Hanstein (TAPS)
25.0
0.0
20.0
R
-0.02
15.0
d /d
10.0 -0.04
5.0 -0.06 0
E =280 MeV E =300 MeV R: p -Data only
0.0 -0.08 R=REM MAMI/TAPS
40.0 Hanstein(DR)
LEGS (E =322 MeV) -0.1
35.0 250 300 350 400
E / MeV
/ ( b/sr)
30.0
25.0 3/2 3/2
Fig. 12. The energy dependence of the ratio E1+ /M1+ is
20.0 shown as solid diamonds. In addition, the energy dependence
d /d
10.0 This is the crucial point of our analysis [66]. This method
oers the advantage of being independent of absolute nor-
5.0
malization and insensitive to many systematic errors, be-
E =360 MeV E =380 MeV
0.0 cause REM is extracted from the ratio of the coe cients
0 40 80 120 160 0 40 80 120 160 C and A tted to the angular distribution of d /d.
CMS CMS
0 (Grad) 0 (Grad) Further, the following identity can be derived [64]:
Fig. 11. Dierential cross sections in the -resonance region.
MAMI/TAPS results are shown with statistical (1 2 %) and 1 C 1 C
A + ( = 90 )
systematic errors (solid circles, this work ref. [64], open dia- R= =
REM , (17)
12 A 12 1 ( = 90 )
monds ref. [59] and crosses ref. [60]).
which depends only on the shape (C/A) of the dieren-
tial cross section d/d and the photon asymmetry
s- and p-wave approximation by the parameterization at CM S = 90 . Using eq. (17), the ratio REM can be
extracted [64]:
dj () q
= (Aj + Bj cos() + Cj cos2 ()), (11) REM = (2.4 0.16stat. 0.24sys ).% (18)
d k
3/2
where q and k denote the center-of-mass momenta of the According to the Fermi-Watson theorem the E1+ and
3/2
pion and the photon, respectively, and j indicates the par- M1+ partial waves have the same phase 33 and the ratio
allel (), perpendicular () and unpolarized (0) compo- 3/2 3/2
E1+ /M1+ is a real quantity. As shown in g. 12, this ratio
nents. The coe cients Aj , Bj and Cj are quadratic or is strongly dependent on the photon energy and varies
bilinear functions of the s- and p-wave amplitudes. In par- from 8% at E = 270MeV to +2% at E = 420MeV.
ticular, d /d is sensitive to the E1+ amplitude, because
of interference with M1+ in the terms
6 Future plans
A = | E0+ |2 + | 3E1+ M1+ + M1 |2 , (12)
B = 2 Re[E0+ (3E1+ + M1+ M1 ) ], (13) The rst round of experiments with the Crystal Ball is
C = 12 Re[E1+ (M1+ M1 ) ]. (14) centered on the rst measurement of the magnetic dipole
moment of the + (1232)-resonance. The magnetic dipole
moment, b , provides us with a simple way for testing the
Furthermore, the ratio validity of the theoretical hadron description in the non-
perturbative sector of QCD. This includes quark soliton
1 C Re(E1+ (M1+ M1 ) ) models, the standard quark models, various eective La-
R= = (15)
12 A | E0+ |2 + | 3E1+ + M1+ M1 |2 grangians and lattice QCD calculations. Our experimental
R. Beck: Experiments with photons at MAMI 181
Entries 4428
ton beams. A unique frozen spin target lled with 1 H, Mean 203.4
Counts
and targets; ii) measurements of the N (1535) magnetic Fig. 14. Invariant mass of two-cluster events for beam photons
dipole moment using p p; iii) a new measurement with energy above 700 MeV after requiring the missing mass to
of the mass. be equal to the mass of proton. The peaks are due to 0 2
and 2 decays.
The new experimental apparatus is shown in g. 13.
The Crystal Ball with TAPS as the forward wall will be
used for detection of photons and nucleons. In addition the spectrometer are measured with an energy resolution
the polar and azimuthal angles of the outgoing proton for
lab > 20 will be measured by the central tracker which is E /E 1.7%/(E (GeV))0.4 ;
based on the DAPHNE cylindrical multiwire proportional
chamber. The chamber will be inserted into the Crystal the angular resolution for photon showers at energies of
Ball beam cavity. 0.05 0.5 GeV is = 2 3 in the polar angle and =
The Crystal Ball was build at SLAC and used in 2 / sin in the azimuthal angle.
J/ measurements at SPEAR and b-quark physics at High granularity and a large acceptance make the
DESY [71]. The CB is constructed of 672 optically iso- Crystal Ball a unique instrument for measuring reactions
lated NaI(Tl) crystals, 15.7 radiation lengths thick. The with multiphoton nal states. The CB detects neutrons
counters are arranged in a spherical shell with an inner ra- with an e ciency of 35% at En = 150 MeV [72].
dius of 25.3 cm and an outer radius of 66.0 cm. The hygro- The rst production run of the CB@MAMI program,
scopic NaI is housed in two hermetically sealed evacuated a measurement of the photon asymmetry in 0 photopro-
hemispheres. Each crystal is shaped like a truncated trian- duction at threshold, was accomplished in July-August
gular pyramid, 40.6 cm high, pointing towards the center 2004. In October 2004 we have started a 600 hours long
of the Ball. The sides on the inner end are 5.1 cm long production run for the measurements of the + (1232)
and 12.7 cm on the far end. Electromagnetic showers in magnetic dipole moment. Figures 14 17 illustrate the
182 The European Physical Journal A
7 0
p ( p)
total ( p p), arb. units
6
0.2
5 0.18
0.16
4
0.14
3 0.12
0.1
2 0.08
0.06
1
0.04
0 0.02
620 640 660 680 700 720 740 760 780 800 820
ELab
, MeV 0
-1 -0.5 0 0.5 1
Fig. 15. The total cross section of p p in arbitrary units cos(CM
)
from 2 decay modes (solid circles) is compared to the
p 3 0 p total cross section (open circles). Above the Fig. 17. Preliminary results for the photon asymmetry times
threshold the p 3 0 p is mainly from 3 0 decay. The beam polarization shown as a function of cos 0 in c. m. for
total cross sections are relatively normalized at E = 730 MeV. E = 375 MeV (triangles), 405 MeV (squares), and 436 MeV
(circles), compared to MAID predictions. The MAID curves
0 are normalized to the experimental data at cos = 0.
p ( p)
0.2
the mass of the proton. Figure 15 shows an excitation func-
0.15 tion for p ()p in arbitrary units. The total cross
section for p 3 0 p is shown on the same gure for com-
0.1 parison. Below threshold 3 0 events are produced via se-
quential decay of resonances, while above the -threshold
0.05 most of the events are produced by the 3 0 decay.
The experimental setup made up of the Crystal Ball
-0 and TAPS is almost perfectly -symmetric. Together with
the good quality polarized MAMI beam it allows high
-0.05 statistics, low systematics uncertainty, polarization mea-
surements. Figure 16 shows the -dependence of the beam
-0.1 photon asymmetry, (), for the reaction p 0 p.
The asymmetry is not corrected for the beam polariza-
-0.15 tion. The data represent about 5 % of the statistics ob-
tained in the course of our most recent ((1232)) run.
-0.2 The photon asymmetry as a function of cos 0 is shown
-150 -100 -50 0 50 100 150
0
( ), deg
in g. 17 for beam photon energies of 375 MeV, 405 MeV,
and 436 MeV in comparison with MAID [73]. The results
Fig. 16. -dependence of the beam photon asymmetry times are very preliminary. The measured distributions are not
polarization, p (), for the reaction p 0 p, where corrected for the beam photon polarization therefore the
() ()
() = ()+ ()
. The function shows clear cos(2) behav- MAID curves are normalized to the data at cos = 0.
ior over the full angular range. The data covers the beam en- The photon asymmetries obtained in our experiment show
ergy interval of 360 450 MeV and is integrated over 0 . good agreement with the MAID evaluations for cos > 0
and slightly deviate in the backward angles. The dierence
quality of the data, showing some characteristic distribu- between our data and MAID gets larger at higher beam
tions. energies.
The invariant mass of two photons for an incident
beam with energy above 700 MeV is shown in g. 14. The I would like to thank the organizers Hartmuth Arenhovel, Hart-
two peaks of the spectra are due to the reactions p mut Backe, Dieter Drechsel, Karl-Heinz Kaiser and Thomas
0 ()p and p ()p. The two-gamma invariant Walcher of the symposium 20 Years of Physics at the Mainz
mass is shown for events with the missing mass equal to Microtron MAMI .
R. Beck: Experiments with photons at MAMI 183
References 37. V. Bernard, N. Kaiser, U.-G. Mei ner, Z. Phys. C 70, 483
(1996).
1. I. Anthony et al., Nucl. Instrum. Methods A 301, 230 38. A. Schmidt, Phys. Rev. Lett. 87, 232501 (2001).
(1991). 39. A. Schmidt, Doktorarbeit, University Mainz (2001).
2. S.J. Hall et al., Nucl. Instrum. Methods A 368, 698 (1996). 40. H. Herminghaus, K.H. Kaiser, H. Euteneuer, Nucl. In-
3. URL: http://wwwa2.kph.uni-mainz.de/A2. strum. Methods A 138, 1 (1976).
4. R. Novotny, IEEE Trans. Nucl. Sci. 38, 379 (1991). 41. D. Lohmann, J. Peise et al., Nucl. Instrum. Methods A
5. G. Audit et al., Nucl. Instrum. Methods A 301, 473 (1991). 343, 494 (1994).
6. A. Partridge et al., Paper presented to the IEEE Meeting, 42. A. Schmidt, Diplomarbeit, University Mainz (1995).
San Francisco, October 1977 (Stanford Linear Accelerator 43. R. Novotny, IEEE Trans. Nucl. Sci. 43, 1260 (1996).
Center, Stanford, 1977). 44. O. Hanstein, D. Drechsel, L. Tiator, Phys. Lett. B 399, 13
7. V.A. Petrunkin, Sov. Phys. JETP 13, 808 (1961). (1997).
8. V.A. Petrunkin, Sov. J. Part. Nucl. 12, 278 (1981). 45. V. Bernard, N. Kaiser, U.-G. Mei ner, Phys. Lett. B 378,
9. V. Olmos de Leon, PhD Thesis, Mainz University (2000). 337 (1996).
10. V. Olmos de Leon et al., Eur. Phys. J. A 10, 207 (2001). 46. V. Bernard, N. Kaiser, U.-G. Mei ner, Eur. Phys. J. A 11,
11. A. Lvov et al., Sov. J. Nucl. Phys. 34, 597 (1981). 209 (2001).
12. A. Zieger et al., Phys. Lett. B 278, 34 (1992). 47. B. Pasquini, D. Drechsel, L. Tiator, Eur. Phys. J. A 23,
13. J.F. Donoghue, B.R. Holstein, Phys. Rev. D 40, 2378 279 (2005).
(1989); J. Bijnens, F. Cornet, Nucl. Phys. B 296, 557 48. V. Bernard, B. Kubis, U.-G. Mei ner, arXiv:nucl-th/
(1988); B.R. Holstein, Comments Nucl. Part. Phys. 19, 0506023v1 (2005).
221 (1990); S. Belluci, J. Gasser, M.E. Sainio, Nucl. Phys. 49. A. Rujula, H. Georgi, S.L. Glashow, Phys. Rev. D 12, 147
B 423, 80 (1994). (1975).
14. U. Burgi, Nucl. Phys. B 79, 392 (1997). 50. C. Becchi, G. Morpurgo, Phys. Lett. 17, 352 (1965).
15. A.N. Ivanov, M. Nagu, N.I. Troitskaya, Mod. Phys. Lett. 51. R. Koniuk, N. Isgur, Phys. Rev. D 21, 1868 (1980).
A 7, 1997 (1992). 52. S.S. Gershteyn et al., Sov. J. Nucl. Phys. 34, 870 (1981).
16. L.V. Filkov, I. Guiasu, E.E. Radescu, Phys. Rev. D 26, 53. D. Drechsel, M.M. Giannini, Phys. Lett. B 143, 329
3146 (1982). (1984).
17. L.V. Filkov, V.L. Kashevarov, Eur. Phys. J. A 5, 285 54. S. Capstick, Phys. Rev. D 46, 2864 (1992).
(1999). 55. A. Wirzba, W. Weise, Phys. Lett. B 188, 6 (1987).
18. V. Bernard, B. Hiller, W. Weise, Phys. Lett. B 205, 16 56. K. Bermuth et al., Phys. Rev. D 37, 89 (1988).
(1988). 57. D.B. Leinweber, Proceedings of the International Confer-
19. M.A. Ivanov, T. Mizutani, Phys. Rev. D 45, 1580 (1992). ence Baryons92 (1992) p. 29.
20. Yu.M. Antipov et al., Phys. Lett. B 121, 445 (1983). 58. A. Buchmann et al., Phys. Rev. C 55, 448 (1997).
21. T.A. Aybergenov et al., Sov. Phys. Lebedev Inst. Rep. 6, 59. R. Beck, H.-P. Krahn et al., Phys. Rev. Lett. 78, 606
32 (1984); Czech. J. Phys. B 36, 948 (1986). (1997).
22. J. Portoles, M.R. Pennington, The Second DAN E 60. G.S. Blanpied et al., Phys. Rev. Lett. 79, 4337 (1997).
Physics Handbook, Vol. 2 (1995) p. 579, hep-ph/9407295. 61. R.M. Davidson, N.C. Mukhopadhyay, Phys. Rev. Lett. 79,
23. J. Boyer et al., Phys. Rev. D 42, 1350 (1990). 4509 (1997).
24. F. Donoghue, B. Holstein, Phys. Rev. D 48, 137 (1993). 62. G. Keaton, R.L. Workman, Phys. Rev. Lett. 79, 4511
25. T. Walcher, Prog. Part. Nucl. Phys. 24, 189 (1990). (1997).
26. J. Ahrens et al., Nucl. Phys. News 4, 5 (1994). 63. O. Hanstein et al., Phys. Lett. B 385, 45 (1996).
27. P. Grabmayer et al., Nucl. Instrum. Methods A 402, 85 64. R. Leukel, PhD Thesis, University Mainz, 2001.
(1998). 65. O. Hanstein et al., Nucl. Phys. A 632, 561 (1998).
28. Ch. Unkmeir, PhD Thesis, Mainz University (2000). 66. R. Beck et al., Phys. Rev. C 61, 035204 (2000).
29. J. Ahrens et al., Eur. Phys. J. A 23, 113 (2005). 67. B. Nefkens et al., Phys. Rev. D 18, 3911 (1978).
30. E. Mazzucato et al., Phys. Rev. Lett. 57, 3144 (1986). 68. M. Kotulla et al., Phys. Rev. Lett. 89, 272001 (2002).
31. R. Beck et al., Phys. Rev. Lett. 65, 1841 (1990). 69. D. Drechsel, M. Vanderhaeghen, Phys. Rev. C 64, 065202
32. M. Fuchs et al., Phys. Lett. B 368, 20 (1996). (2001).
33. J.C. Bergstrom et al., Phys. Rev. C 53, R1052 (1996); 55, 70. A.I. Machavariani, A. Faessler, arXiv:nucl-th/0202060
2016 (1997). (2002).
34. P. de Baenst, Nucl. Phys. B 24, 633 (1970). 71. E.D. Bloom, C.W. Peck, Annu. Rev. Nucl. Sci. 33, 143
35. I.A. Vainshtein, V.I. Zakharov, Nucl. Phys. B 36, 589 (1983).
(1972). 72. T.D. Stanislaus et al., Nucl. Instrum. Methods A 462, 12
36. V. Bernard, J. Gasser, N. Kaiser, U.-G. Mei ner, Phys. (2001).
Lett. B 268, 291 (1991). 73. D. Drechsel et al., Nucl. Phys. A 645, 145 (1999).
Eur. Phys. J. A 28, s01, 185195 (2006)
DOI: 10.1140/epja/i2006-09-019-0 EPJ A direct
electronic only
/
Published online: 31 May 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Coherent radiation in the range from soft X-rays up to hard X-rays, produced by the low-
emittance electron beam of MAMI, can be used for various applications. Novel types of interferometers
have been developed for the measurement of the complex index of refraction of thin self-supporting foils.
For the vacuum ultraviolet and soft X-ray region the interferometer consists of two collinear undulators,
and a grating spectrometer. A foil placed between the undulators causes a phase shift and an attenuation of
the oscillation amplitude. The complex index of refraction has been measured at the L2,3 -absorption edges
of nickel. A novel method is described for the measurement of the X-ray magnetic circular birefringence.
For the hard X-ray region the interferometer consists of two foils at which the 855 MeV electron beam
produces transition radiation. Distinct interference oscillations have been observed as a function of both,
the photon emission angle and the distance between the foils. The refractive index decrement () of a
2 m thick nickel sample foil has been measured at X-ray energies around the K absorption edge at 8333 eV
and at 9930 eV with an accuracy of better than 1.5 %. The line width of parametric X radiation (PXR) was
measured in backward geometry with a Si single-crystal monochromator. Upper limits of the line width of
42 meV, 50 meV, and 44 meV, have been determined for the (333), (444) and (555) reections at photon
energies of 5932 eV, 7909 eV, and 9887 eV, respectively. Small angle scattering of the electrons in the crystal
leads to a stochastic frequency modulation of the exponentially damped wave train which results in the
line broadening. To elucidate the quest if the production of PXR is a kinematical or a dynamical process
the radiation from silicon single-crystal targets, emitted close to the electron direction, has been studied.
The observed interference structures and the narrow-band radiation in forward direction shows that PXR
is produced in a dynamical process.
PACS. 07.60.Ly Interferometers 78.20.Ci Optical constants (including refractive index, complex di-
electric constant, absorption, reection and transmission coecients, emissivity) 41.60.-m Radiation by
moving charges 41.50.+h X-ray beams and X-ray optics
Undulator Radiation Transition Radiation Parametric production and monochromatisation of the radiation take
X-Radiation (PXR) place in the same crystal. The expected small spectral
line width of PXR would promise an abundance of ap-
N S N S
plication possibilities, e.g. within the eld of solid-state
physics. However, line broadening by multiple scattering
S N S N of the electrons in the crystal may spoil the superb line
width. In sect. 3 our experiments addressing this question
Channeling Radiation Smith-Purcell are reviewed, including fundamental aspects like the ques-
(CR) Radiation (SP) tion whether the process of PXR production is of kinemat-
ical or of dynamical nature.
The paper closes with a conclusion and an outlook.
500
d '
Variable distance 0
1000
with sample =Z = 854.3 eV
Fig. 2. Interferometry with spatially separated coherent X-ray
counts
emitters. 800
600
as a Fourier analyser of the wave trains. The two result- 400
ing plane waves have a phase dierence of = v (, d) 200
(v is the velocity of the electron) and interfere in the de-
tector, resulting in oscillations of the intensity I(d), if the 0
distance d is varied. A sample foil placed between the two 2000 with sample =Z = 874.1 eV
sources produces an additional phase shift and attenuation 1500
of wave 2. Consequently, both quantities, i.e. the refrac-
tive index decrement and the absorption index , can 1000
be extracted from the measured interference oscillations
I(d) with and without the foil between the sources. This 500
holds independently of the nature of the emission process,
provided that the produced X-rays remain coherent. 0
210 220 230 240 250 260
d [ mm ]
2.1 The soft X-ray interferometer
Fig. 3. Intensity oscillations as a function of the distance d
For photon energies in the range of about 100 eV up to between the undulators with and without the self-supporting
2 keV we use two identical undulators (period length LU = 83.2 g/cm2 nickel sample foil at three dierent photon ener-
gies at the L2,3 absorption edges. Note the change of sign of
12 mm, number of periods 10, undulator parameter K =
the phase shift at the photon energy of 854.3 eV.
1.1) and a grating spectrometer. The recorded intensity
with a foil between the undulators is given by
I(d) = |A1 |2 + |A2 |2 e2 c ()t0 + 2|A1 | |A2 |e c ()t0 resolution of 0.44 eV. Typical measured intensity oscilla-
tions are shown in g. 3. The extracted optical constants
K2
cos (, d) + ()t0 + 2 LU (1) at the Ni L2,3 are shown in g. 4. A high accuracy has
c 4
been reached, even in the region where in which the
with A2 being the amplitude of the upstream undulator, optical constants can be determined from reectivity mea-
A1 that of the downstream one and t0 the thickness of the surements only with large uncertainties [20]. The remark-
foil. able fact about this measurement is the hight of the L3
The interferometer has been developed at the Mainz resonance with its maximum value = (5.640.24)103 .
Microtron MAMI and its performance was demonstrated This value corresponds to an imaginary scattering factor
with measurements at the K absorption edge of carbon f = 65.7 2.8. In ref. [21] a mass attenuation coecient
at 284 eV. Details of this experiment can be found else- = 2 104 cm2 /g was determined from which, with the
where [18]. The visibility (coherence), dened by C = relation = (4), an f = 24 can be calculated with an
(Imax Imin )/(Imax + Imin ) without sample foil, is close estimated uncertainty of 10 %. The dierence may orig-
to its maximum value C = 1. No loss of coherence was inate from the better resolution in our experiment. Cor-
observed over the scanning distance of 15 cm. Therefore, related with this L3 absorption resonance is a change of
the optical constants and could be extracted by a t sign of the refractive index decrement (). It is interest-
with simple cosine functions. ing to notice that for () < 0 the real part of the refrac-
Measurements were also performed at the L2 - tive index 1 () is larger than 1 and a monochromatic
absorption and L3 -absorption edges of nickel at 871 eV Cherenkov radiation with an energy of 855 eV will be emit-
and 855 eV, respectively. The experimental setup was sim- ted. This fact has been pointed out in the literature.
ilar with that described in ref. [18] with the following The strong absorbtion line at the L3 edge of Ni is
modications: the third harmonics of the undulator was the result of an allowed dipole transition between the
chosen as radiation, which was generated at an electron 2p3/2 core state and empty 3d valence states above the
energy of 766.3 MeV. The radiation was analyzed with a Fermi energy. This transition exhibits a strong X-ray
variable line spacing (VLS) grating [19] with an energy Magnetic Circular Dichroisim (XMCD) eect which can
188 The European Physical Journal A
6
+/- 90
4
[ 10 ]
e-
-3
Undulator 1
2 Movable Gap
E
0 Beam tube
L3 L2 Undulator 2
2 3 mm
1m
0
[ 10 ]
-3
60 447
335
40
row
223
20
111
0 0
0 20 40 60 80 100 120 140 160 180 200
col mn
Fig. 8. Measured interference pattern at a xed foil distance
d = 475.0(2) m. The central photon energy h = 9929 eV is
well above the K absorption edge of nickel. Grey levels indicate
the intensities.
1000
Target Crystal Analyzer
(111)
(T1 = 145 K) Crystal (T2 = 296 K)
(333)
1, x 2 Slit
(444)
(555)
2, x 2
H1 H2 1.12m 10
855 MeV Si-Drift
Electron Beam Detector
15.3m
1
0.1
3.2 Measurement of forward-diracted PXR
0.0
-0.4 -0.3 -0.2 -0.1 0.0 Up to now it could not be decided experimentally whether
'=Z [eV] PXR emission is a kinematical or a dynamical process.
The reason has been discussed by Nitta in a recent pa-
Fig. 12. Line shape measurements of PXR at 1,x /2 = 5 mrad per [61]. He showed that the rst-order approximation
for dierent reections as indicated by (nnn) values; h = of the dynamical calculation gives the kinematical ex-
h h0 , with h0 the energy of the DTR reection. pression. Extremely accurate absolute intensity measure-
ments would be required to gure out a dierence. Bary-
shevsky [62] proposed to search for the predicted forward-
in g. 12 by the full lines. The resulting PXR widths are diracted wave (FDPXR) which is associated to PXR and
shown in g. 13 together with the natural line width ac- emitted close to the direction of the electron propagation.
cording to eq. (2) and the width of the Darwin-Prins curve. Similar proposals have also been communicated by Na-
Only for the (333) reex the observed line width is smaller sonov [63,64]. In ref. [65] the observation of narrow FD-
than the width of the Darwin-Prins curve. It could not be PXR structures from a 410 m thick tungsten single crys-
excluded that imperfections of the analyzer crystal itself tal at photon energies of 28.3 and 40 keV is reported.
broaden the higher reections. Therefore, the real PXR At the Mainz Microtron MAMI experiments were per-
line could be somewhat smaller as the t results shown in formed for the search of the forward-diracted wave (FD-
g. 13. PXR) in single silicon crystals of various thicknesses [66].
192 The European Physical Journal A
x 0
z
HA
x Column Number Column Number
O
0 0
Fig. 16. Results of simulation at photon energy of 10.554 keV
1000
5000 and target thickness of 58 m (left) and 1000 m (right). Shown
are the number of photons N per pixel and electron for one
Events
500
row of the pn CCD detector as a function of the column num-
ber. From the upper to the lower panel the rotation angle of the
0 0
target crystal x was varied in steps of x = 0.5894 mrad.
1000
5000 The beam spot sizes of the experiment and scattering of the
electron beam were taken into account. The residual interfer-
500
ence oscillations originate from the interference of the remain-
ing 4% amplitude created at the entrance interface with the
0 0
amplitude at the exit interface of the crystal. However, these
0 50 100 150 200 0 50 100 150 200
oscillations are smoothed out in a real experiment in which
Column Column summation is made over several or all rows of the pn CCD
detector.
Fig. 15. Measurements at photon energy of 10.554 keV and
target thickness of 58 m (left) and 1000 m (right). Shown
are intensity distributions summed over all 64 rows of the
pn CCD detector as a function of the column number. From
observation angle x . Since its reecting power of the crys-
the upper to the lower panel the rotation angle x of the tal monochromator exhibits energetically a narrow-band
target crystal was varied in steps x = 0.5894 mrad. Beam characteristics, quasi-monochromatic intensity structures
current: 53.5 nA, exposure time: 600 s. Left panel: Beam spot emitted from the target crystal can be detected by this
size about 500 m (FWHM) horizontally and 434 m (FWHM) experimental arrangement.
vertically. The destructive interference fringes can clearly be Experiments were performed with target crystals of
recognized. Right panel, upper curves: Beam spot size about varying thickness and for dierent photon energies. As an
500 m (FWHM) horizontally and 434 m (FWHM) vertically. example the intensity distributions of the experiment with
Right panel, lower curves: Reduced beam spot size 114 m 58 m and 1 mm crystal thicknesses are shown in g. 15.
(FWHM) horizontally and 200 m (FWHM) vertically. The most striking features are the structures which move
across the pn CCD detector if the rotation angle x of the
target crystal around the vertical y-axis is varied. These
The basic idea of the experiment will be explained by structures are for the thin targets interferences of the ra-
means of g. 14. A silicon single-crystal target was po- diation amplitudes created at the entrance and exit inter-
sitioned in such a way that the PXR reex at a photon faces of the crystal which originate from a resonance in
energy h0 = 10.554 keV is located at twice the Bragg the dispersion surface of the electron in the crystal. The
angle 0 = 10.797 in the horizontal plane of drawing. pronounced peak structures observed for the thick target
The radiation in forward direction close to the electron which are clearly correlated to the interference structures
direction was analyzed with a at silicon single-crystal of the thin targets are interpreted as FDPXR contribu-
monochromator in combination with a pn CCD camera as tions to the smooth transition radiation background from
a position-sensitive and energy-resolving photon detector. the downstream interface of the target crystal.
The quasi-monochromatic FDPXR peak energy matches The interference structures can quantitatively be ex-
with the energy of the analyzer crystal at only one specic plained in the framework of the well-known TR production
W. Lauth et al.: Coherent X-rays at MAMI 193
mechanism utilizing a generalized formation length for duction of PXR. For thin crystals pronounced interference
crystalline matter [66]. Within this model, which is based structures in forward direction have been observed when
on the formalism described in ref. [67], the resonance is about a Bragg condition was fullled. This interference is
connected to forward-diracted PXR (FDPXR). Calcu- corroborated by the narrow FDPXR lines observed with a
lated emission spectra of this model, including multiple thick Si crystal for which essentially only the exit interface
scattering of the electrons, are displayed in g. 16. The of the target crystal contributes to the observed intensity.
simulations are in good agreement with the experimental Our experimental results show that PXR production is a
observation. dynamical rather than a kinematical process.
O. Kettig, G. Kube, W. Lauth, H. Schope, Th. Walcher, 64. N. Nasonov, V. Sergienko, N. Noskov, Nucl. Instrum.
Phys. Rev. Lett. 79, 2462 (1997). Methods Phys. Res. B 201, 67 (2003).
60. H. Backe, C. Ay, N. Clawiter, Th. Doerk, M. El-Ghazaly, 65. N. Aleinik, A.N. Baldin, E.A. Bogomazova, I.E. Vnukov,
K.-H. Kayser, O. Kettig, G. Kube, F. Hagenbuck, W. B.N. Kalinin, A.S. Kubankin, N.N. Nasonov, G.A. Nau-
Lauth, A. Rueda, A. Scharafutdinov, D. Schro, T. Weber menko, A.P. Potylitsyn, A.F. Sharafutdinov, JETP Lett.
in: W. Greiner, A. Solovyov, S. Misicu (Editors) Proceed- 80, 393 (2004) (Pisma Zh. Eksp. Teor. Fiz. 80, 447
ings Symp. Channeling - Bent Crystals - Radiation Pro- (2004)).
cesses, Frankfurt (Germany) 2003 (EP Systema, Debre- 66. H. Backe, A. Rueda, W. Lauth, N. Clawiter, M. El-
cen, 2003) p. 41. Ghazaly, P. Kunz, T. Weber, Nucl. Instrum. Methods
61. H. Nitta, J. Phys. Soc. Jpn. 69, 3462 (2000). Phys. Res. B 234, 138 (2005).
62. V.G. Baryshevsky, Nucl. Instrum. Methods Phys. Res. B 67. A. Caticha, Phys. Rev. A 40, 4322 (1989).
122, 13 (1997). 68. C.D. Back, D. Weller, J. Heidmann, D. Mauri, D. Guarisco,
63. A. Kubankin, N. Nasonov, V. Sergienko, I. Vnukov, Nucl. E.L. Garwin, H.C. Siegmann, Phys. Rev. Lett. 81, 3251
Instrum. Methods Phys. Res. B 201, 97 (2003). (1998).
Eur. Phys. J. A 28, s01, 197 208 (2006)
DOI: 10.1140/epja/i2006-09-021-6 EPJ A direct
electronic only
/
Published online: 6 June 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. Experiments have been performed to explore the potential of the low emittance 855 MeV electron
beam of the Mainz Microtron MAMI for imaging with coherent X-rays. Transition radiation from a micro-
focused electron beam traversing a foil stack served as X-ray source with good transverse coherence.
Refraction contrast radiographs of low absorbing materials, in particular polymer strings with diameters
between 30 and 450 m, were taken with a polychromatic transition radiation X-ray source with a spectral
distribution in the energy range between 8 and about 40 keV. The electron beam spot size had standard
deviation h = (8.6 0.1) m in the horizontal and v = (7.5 0.1) m in the vertical direction. X-ray
lms were used as detectors. The source-to-detector distance amounted to 11.4 m. The objects were placed
in a distance of up to 6 m from the X-ray lm. Holograms of strings were taken with a beam spot size
v = (0.50 0.05) m in vertical direction, and a monochromatic X-ray beam of 6 keV energy. A good
longitudinal coherence has been obtained by the (111) reection of a at silicon single crystal in Bragg
geometry. It has been demonstrated that a direct exposure CCD chip with a pixel size of 13 13 m 2
provides a highly e cient on-line detector. Contrast images can easily be generated with a complete
elimination of all parasitic background. The on-line capability allows a minimization of the beam spot
size by observing the smallest visible interference fringe spacings or the number of visible fringes. It has
been demonstrated that X-ray lms are also very useful detectors. The main advantage in comparison
with the direct exposure CCD chip is the resolution. For the Structurix D3 (Agfa) X-ray lm the standard
deviation of the resolution was measured to be f = (1.2 0.4) m, which is about a factor of 6 better
than for the direct exposure CCD chip. With the small eective X-ray spot size in vertical direction of
v = (1.20.3) m and a geometrical magnication of up to 7.4 high-quality holograms of tiny transparent
strings were taken in which the holographic information is contained in up to 18 interference fringes.
PACS. 87.59.Bh X-ray radiography 52.59.Px Hard X-ray sources 41.50.+h X-ray beams and X-ray
optics 07.85.Fv X- and gamma-ray sources, mirrors, gratings, and detectors 07.85.Nc X-ray and
gamma-ray spectrometers
1 Introduction particular for low-Z materials, the phase shift for X-rays
is higher than the absorption of the incident X-rays. Also,
The contrast in conventional absorption X-ray imaging is for the radiography based on the phase shift mechanism,
based on the dierence in the absorption of the materials the absorbed dose is considerably lower in comparison to
constituting the sample. Thin samples of light elements, the conventional absorption radiography, see, e.g., refs. [1,
such as soft tissues and organic materials with Z 8, show 2,3].
a weak absorption contrast even at low X-ray energies, i.e.,
X-ray phase contrast imaging can be carried out with
the big deciency is that the conventional absorption ra-
various methods, for an overview see the recent ref. [4]. In
diography cannot distinguish between materials with simi-
particular, it has been pointed out by Wilkins et al. [5]
lar attenuation coe cients. For low-Z materials, however,
that a very simple experimental setup with a polychro-
a high contrast could be obtained if the phase shift of
matic X-ray source of good transverse coherence, i.e. a
the X-rays introduced by the object could be exploited
small micro-focused spot, is already su cient. Informa-
instead of the intensity of the transmitted wave. The en-
tion can be supplied by such a method on the sample
hancement of the contrast is attributed to the fact that, in
morphology, i.e. its boundaries, interfaces and location of
a
Former PhD Scholarship Holder in the Long Term Mission small features, see e.g. ref. [6,7,8]. If, in addition, the X-
System from the Arabic Republic of Egypt. ray source emits monochromatic X-rays, holograms can
b
Present address: Deutsches Elektronen-Synchrotron be taken. The experimental setup is similar to that of
DESY, Notkestra e 85, D-22603 Hamburg, Germany. Gabor in-line holography [9]. In principle, such a setup
198 The European Physical Journal A
]
Object Intensity
,PD[
( ; [ , ] ), 0 ,PLQ
Detector
S
Source
[VR [RG
-3
-
10 Si (111) crystal
Polyimide NH9
monochromator
foil stack
0
H 10 20 30 40 50
[keV] 10 4
H to
beam dump
600 MeV e- beam
Quadrupole
Transition radiation 38.5 6 keV
doublet
Micro focus
for a moderate detector resolution if, e.g., CCD-chips in a with a large dynamic range and linear relationship be-
direct exposure mode are used, see below. tween the incident radiation intensity and the response of
For the preparation of a micro-focused electron beam, the detector. Such conditions can be fullled by a charge-
a low beam emittance in horizontal and vertical directions coupled device (CCD) or an X-ray lm. For the current ex-
is of particular importance. The emittance of the MAMI periments the CCD system ANDOR DO-434 BN CCD [22]
electron beam in horizonal direction is bigger than the was used. It contains a back-illuminated CCD low-noise
emittance in vertical direction because the electrons emit sensor from Marconi CCD47-10 [23] with 1024 1024 pix-
synchrotron radiation in the bending magnets of race- els of size 13 13 m2 . The chip has a good quantum
track microtron 3. The horizontal emittance grows rapidly e ciency over a wide spectral range. For X-rays of 6 keV
above an electron beam energy of 400 MeV, while the ver- energy it amounts to still about 45%. These features oer
tical emittance still decreases. As a compromise, a beam the opportunity to use the CCD chip in the direct ex-
energy of 600 MeV was chosen for which the emittances posure mode in which the signal is generated by direct
are h = 2.3 m mrad and v = 0.52 m mrad in the hori- energy deposition of X-rays in the sensitive layer of ap-
zontal and vertical directions, respectively. proximately 10 m thickness.
The polyimide foil stack to produce transition radia- Direct-exposure CCD camera chips have, compared
tion is optimized for a high X-ray ux at a photon energy with X-ray lms, the big advantage that they have a good
of 6 keV at the electron beam energy of 600 MeV. The cal- linearity over a wide dynamical range, a good signal-to-
culated photon energy spectrum is shown in the inset of noise ratio, and that they are on-line capable. The latter
g. 9. The at silicon single crystal with its surface par- fact is very important since contrast or normalized con-
allel to the (111) crystal plane acts as a mirror for the trast images can easily be generated in which all parasitic
TR photons. However, the mirror is energy dispersive in background, originating not from the object, can be elim-
the horizontal direction. The deviation of the photon inated. The disadvantage of a moderate spatial resolution
energy, dened by the equation h = hB (1 + ), from in comparison to an X-ray lm can be alleviated by a ge-
the nominal Bragg energy ometrical magnication. In reality, however, the spatial
resolution is larger than the pixel size because of so-called
2 h2 + k 2 + l2 hc split events in which the deposited energy is shared by
hB = (13) neighboring pixels.
a0 2 sin B
The Structurix D3 X-ray lm from Agfa is a useful
is approximately given by the expression [21] detector as well. The main advantage in comparison with
the direct exposure CCD chip is its very good resolution.
(0 ) x The standard deviation of the resolution was measured to
= , (14)
2 sin2 B tan B be f = (1.2 0.4) m, which is about a factor of 6 better
than for the direct-exposure CCD chip. The main disad-
where x = B is the deviation from the nominal vantage of the X-ray lm is the missing on-line capability
Bragg angle B . The integers h, k, l are the Miller indices, with the consequence that the generation of normalized
a0 = 5.4309 A the lattice constant, and (0 ) the real contrast images is rather involved.
part of the dielectric susceptibility 0 . The Bragg angle The procedure to obtain the intensity information from
for hB = 6 keV amounts for the (111) reection to B = the photographic density is similar to that already de-
19.25 . scribed in sect. 3.2. The X-ray lm was digitized with a
In in-line holography a scattered wave from the object lm scanner (Nikon Coolscan LS 4000 [17]) and with an
interferes with an unscattered wave from the source. For optical microscope equipped with a high-resolution 8-bit
an assumed transverse coherence length of LT = 250 m, CCD camera (F-View XS [24]). From this system limita-
at a distance of 13.6 m a x = 18.4 rad results. The tions are expected because the dynamical range cannot
angular spread which originates from the beam spot size be better than the digitization depth of the ADC (1:256),
and the pixel resolution is less than 1 rad and can be ne- while the X-ray lm has a dynamical range which is more
glected. The angle x corresponds, according to eq. (14), than a factor of 10 better (3.5 decades corresponding to
to a relative energy shift of 5.3 105 or h = 0.32 eV. 1:3160). Since, in addition, the illumination time was se-
This means that two waves with slightly dierent ener- lected automatically by the scanner after the part of in-
gies must interfere which is only possible if the longitu- terest of the picture and the optical magnication were
dinal coherence length LL is long enough. The width of selected, the holograms were digitized at various positions
the reecting power ratio of the monochromator crystal of the string. Along the imaged strings the exposure is
is = 1.4 104 corresponding to h = 0.84 eV [21], changing and the sector with the best contrast was se-
and a longitudinal coherence length LL = 0.52 / = lected for further analysis.
0.5/ = 0.74 m results. This value is su ciently large
for all objects investigated in this work which had thick- 4.3 Measurements and discussion
nesses in the sub-mm range, since at a refractive index
4.3.1 Investigation of the transverse coherence in horizontal
decrement of = 1 106 the optical path dierence is
direction
less than 0.01 m.
Hard X-ray holography requires, like refraction con- To study the coherence in the horizontal and vertical di-
trast radiography, a two-dimensional resolving detector rections, radiographs of two polymer strings of the same
M. El-Ghazaly et al.: X-ray phase contrast imaging at MAMI 205
4.4 Conclusions
4. F. Pfeier, T. Weitkamp, O. Bunk, Ch. David, Nature 16. B.L. Henke, J.Y. Uejio, G.F. Stone, C.H. Dittmore, F.G.
Physics advance online publication www.nature.com/ Fujiwara, J. Opt. Soc. Am. B. 11, 1540 (1986).
naturephysics, published online: 26 March 2006; 17. http://www.filmscanner.info/NikonSuperCoolscan-
doi:10.1038/nphys265. 4000ED.html.
5. S.W. Wilkins, T.E. Gureyev, D. Gao, A. Pogany, A.W. 18. Georg Joos, Erwin Schopper, Grundriss der Photogra-
Stevenson, Nature (London) 384, 335 (1996). phie und ihrer Anwendungen besonders in der Atomphysik
6. Xizeng Wu, Hong Liu, Med. Phys. 30, 2169 (2003). (Akademische Verlagsgesellschaft M. B. H., Frankfurt am
7. T. Takeda, A. Momose, E. Ueno, Y. Itai, J. Synchrotron Main, 1958).
Rad. 5, 1133 (1998). 19. Y. Hwu, H.H. Hsieh, M.J. Lu, W.L. Tsai, H.M. Lin, W.C.
8. R.A. Lewis, Phys. Med. Biol. 49, 3573 (2004). Goh, B. Lai, J.H. Je, C.K. Kim, D.Y. Noh, H.S. Youn, G.
9. D. Gabor, Nature 161, 777 (1948). Tromba, G. Margaritondo, J. Appl. Phys. 86, 4613 (1999).
10. P. Spanne, C. Raven, I. Snigireva, A. Snigirev, Phys. Med. 20. O. Chubar, A. Snigirev, S. Kuznetsov, T. Weitkamp, V.
Biol. 44, 741 (1999). Kohn, Proceedings DIPAC 2001, ESRF, Grenoble, France.
11. P. Cloetens, R. Barrett, J. Baruchel, J. Guigay, M. 21. A. Caticha, Phys. Rev. A 40, 4322 (1989).
Schlenker, J. Phys. D 29, 133 (1996). 22. http://www.andor-tech.com/germany/products/oem.
12. Z.W. Hu, B. Lai, Y.S. Chu, Z. Cai, D.C. Mancini, B.R. cfm
Thomas, A.A. Chernov, Phys. Rev. Lett. 87, 148101 23. http://www.data.it/support/data sheets/e2vtech/
(2001). 47-10back.pdf
13. R.W. James, The optical Principles of the Diraction of 24. http://www.olympus.pl/pliki/mikroskopy/dokumenty/
X-rays (Cornell University Press, 1965). LM cameras ENG.pdf.
14. V. Kohn, I. Snigireva, A. Snigirev, Opt. Commun. 198, 25. C. Raven, Microimaging and Tomography with High En-
293 (2001). ergy Coherent Synchrotron X-Rays (Shaker Verlag, 1998).
15. Mahmoud El Ghazaly, X-ray Phase Contrast Imaging at 26. R.W. Gerchberg, W.O. Saxton, Optik 35, 237 (1972).
the Mainz Microtron MAMI, Dissertation, Institut fur
Kernphysik, Universitat Mainz, 2005.
Author index
dHose N.: Virtual Compton Scattering at MAMI 117 Ostrick M.: Electromagnetic form factors of the nucleon
81
El-Ghazaly M., Backe H., Lauth W., Kube G., Kunz P.,
Sharafutdinov A. and Weber T.: X-ray phase contrast Rohe D. (A1 and A3 Collaboration): Experiments with
imaging at MAMI 197 polarized 3 He at MAMI 29
Hammer H.-W.: Nucleon form factors in dispersion theory Scherer S.: Chiral perturbation theory 59
49 Schmieden H.: Photo- and electro-excitation of the
Hillert W.: The Bonn Electron Stretcher Accelerator -resonance at MAMI 91
ELSA: Past and future 139 Schwamb M.: Few-nucleon systems (theory) 39
Sharafutdinov A. El-Ghazaly M.
Jankowiak A.: The Mainz Microtron MAMI Past and Sharafutdinov A. Lauth W.
future 149
Thomas A.: The Gerasimov-Drell-Hearn sum rule at
Kettig O. Lauth W. MAMI 161
Kowalski S.: Parity violation in electron scattering 101
Kube G. El-Ghazaly M. Vanderhaeghen M.: Two-photon physics 71
Kunz P. El-Ghazaly M.
Kunz P. Lauth W. Weber T. El-Ghazaly M.
Weber T. Lauth W.
Lauth W., Backe H., Kettig O., Kunz P., Sharafutdinov
A. and Weber T.: Coherent X-rays at MAMI 185
Lauth W. El-Ghazaly M.
Eur. Phys. J. A 28, s01, 209 219 (2006)
DOI: 10.1140/epja/i2006-09-020-7 EPJ A direct
electronic only
/
Published online: 7 June 2006
c Societa Italiana di Fisica / Springer-Verlag 2006
Abstract. The development over the last twenty years of the physics program and the experimental facilities
at the Mainz Microtron MAMI will be reviewed. Ground-breaking contributions have been made to the
development of experimental techniques and to our understanding of the structure of nucleons and nuclei.
PACS. 29.17.+w Electrostatic, collective, and linear accelerators 25.20.-x Photonuclear reactions
25.30.Bf Elastic electron scattering 25.30.Dh Inelastic electron scattering to specic states
1 Introduction and what are the dierences in quark wave functions be-
tween the ground state and the excited states.
The goal of nuclear physics is to study the properties of Many of the issues mentioned above can be investi-
nuclei, and to understand these properties on the basis of gated using electron scattering. Electrons interact only
the fundamental theory of the constituents making up the with the charged constituents of the object under investi-
nucleus. Quantum-chromodynamics (QCD) has emerged gation. The interaction is described by Quantum-Electro-
as the leading candidate for the theory of hadronic inter- dynamics (QED), and is su ciently small to be handled
actions. Presently, QCD cannot be solved in the strong- with perturbative methods.
coupling regime due to the lack of appropriate pertur- In the past, the usefulness of the electromagnetic probe
bative solution. The best hope in the near future is to was limited by the technical features of the available elec-
use large-scale numerical calculations to approximate the tron accelerators. In particular, coincidence experiments
space-time continuum by a discrete lattice in the frame- and the operation of large acceptance detectors were ham-
work of Lattice QCD (LQCD). Due to limitations in com- pered by the low duty-cycle of the electron beams. Dra-
pute power, the lattice spacing is still fairly coarse, and matic progress in accelerator and detector technology has
the masses of the quarks used are much larger than their made it possible to overcome these limitations and to
actual values. These limitations require signicant extrap- study electromagnetic processes with an accuracy that is
olations which need to be constrained theoretically. no longer limited by technical problems.
For zero-mass quarks, QCD can be solved in a rigor- The Institut fur Kernphysik (Institute for Nuclear
ous way via Chiral Perturbation Theory (PT). Again, an Physics) at the University of Mainz in Germany has been
extrapolation is required, this time from zero-mass quarks at the forefront of this development. For more than 20
up to the actual values. An interesting recent development years, the Institute has developed novel electron accel-
is the use of functional forms derived from PT to extrap- erators (the MAMI series of microtrons) and the corre-
olate LQCD results. sponding experimental equipment, and has used those de-
Before these extrapolations can be trusted, it is very vices for ground-breaking research into the electromag-
important to verify the predictive power of PT via ex- netic structure of nucleons and nuclei.
perimental tests, in particular via pion photoproduction On the occasion of the retirement of six key people
close to threshold. (Hartmuth Arenhovel, Hartmut Backe, Dieter Drechsel,
On the experimental side, valuable information on the Jorg Friedrich, Karl-Heinz Kaiser, and Thomas Walcher)
properties of bound quark systems is still lacking. In par- from the Institute, a symposium was held in October 2005
ticular, the knowledge of the spatial distribution of the to review 20 years of Physics at MAMI and to commem-
charges and the currents inside the nucleon is not satis- orate their contributions.
factory, especially for the neutron. Our knowledge of the This paper will attempt to review the major contri-
excited states of the nucleon is still insu cient, e.g. what butions MAMI and its user community have made to the
are the degrees-of-freedom governing the mass spectrum, eld, to identify the particular circumstances that made
these contributions possible, and to speculate on their last-
a
e-mail: mecking@jlab.org ing impact.
210 The European Physical Journal A
Fig. 1. Layout of the MAMI accelerators and the experimental areas. MAMI A is located in the RTM2 area, MAMI B in
RTM3, the new MAMI C in the area labeled HDSM.
Table 1. MAMI microtron development. eective way to build an accelerator capable of delivering
a high quality electron beam [1]. The layout of the accel-
Year Activity erator and its experimental areas is shown in g. 1. The
1975 Proposal for a Race-Track Microtron microtron design relies on sending the beam repeatedly
(design by H. Herminghaus et al.) through the same room-temperature accelerating struc-
1979 14 MeV beam from MAMI A1 ture with moderate energy gain per turn. Recirculation is
1982 Preliminary Sonderforschungsbereich achieved by two homogeneous 180 end-magnets. The size
(SFB) established of these end-magnets for the last microtron stage, MAMI
1983 183 MeV beam from MAMI A2 B, is evident from g. 2. The perpendicular entry and exit
1983 - 87 MAMI A operation with a total of 18,700 h of the electron orbits at the end-magnets results in simple
1983 - 90 Development of the 855 MeV MAMI B and robust beam optics. Due to the continuous-wave na-
1984 SFB 201 established ture of the radio-frequency power and the constant mag-
1990 First 855 MeV beam from MAMI B netic eld, the quality of the beam is very high: an energy
(rst experiment by A2 Collaboration) spread of E/E = 1.5 105 and an emittance of = 8
1990 - 2005 MAMI B operation with a total of 82,843 h 109 m is achieved routinely. A laser-driven polarized gun
1999 Sonderforschungsbereich 443 established produces electron beams with 80% polarization. Parity vi-
2000 Approval of 1.5 GeV HDSM olation experiments are possible since helicity-correlated
(Harmonic Double-Sided Microtron,
changes in the beam parameters are very small: energy
design by K.-H. Kaiser et al.)
variations of E/E 108 and position variations of x
2001 - 03 Installation of the four HDSM magnets
2006 Commissioning and begin of physics
100 nm have been achieved. Particularly impressive is the
high operational stability: overnight and during weekends,
the MAMI microtrons are routinely operated by students.
2 MAMI microtron development It has been the conventional wisdom in the accelerator
community that the maximum energy of a microtron is
The history of the MAMI microtron development is sum- limited to about one GeV since the construction of the
marized in table 1. The pioneering development of MAMI end-magnets which increase rapidly in size with increasing
A and B at Mainz has established the microtron as a cost- energy becomes technically and nancially impractical.
B.A. Mecking: Twenty years of physics at MAMI What did it mean? 211
Spectrometer A B C
conguration QSDD D QSDD
pmax [MeV/c] 665 810 490
[msr] 28 5.6 28
min 18 7 18
p/p [%] 20 15 25
LINAC I (4.90GHz)
n 1500MeV
Extractio
43 recirculations
In
je
ct
io
The design and construction of the 1.5 GeV Harmonic 3 Experimental equipment at MAMI
Double-Sided Microtron (HDSM, design by K.-H. Kaiser
et al.) [2] is poised to shatter that boundary. Building The broad physics program at MAMI requires an equally
on the experience with the previous microtrons, the chal- broad range of experimental equipment, from high-resolu-
lenging HDSM design relies on two parallel accelerating tion magnetic spectrometers to large acceptance detectors.
sections joined by four inhomogeneous 90 end-magnets Most of the electron scattering instrumentation has been
with a weight of 250 metric tons each (see g. 3 for a lay- provided by the Institute, a large fraction of the equip-
out). The strong vertical defocusing at the entrance and ment for the tagged photon experiments has been con-
exit of the magnets is compensated by a radial gradient tributed by the user community. For electron scattering
eld. Meeting the microtron coherence condition within experiments, the three-spectrometer system oers an un-
the conned space of the existing experimental area forces precedented combination of momentum resolution, solid
the fundamental accelerating frequency to be twice the angles, and momentum range. The parameters are given
frequency of the MAMI B microtron. Phase stability con- in table 2, a photograph is shown in g. 4.
siderations require to leave one of the two accelerating For experiments with real photons, the Glasgow-Mainz
sections at the present MAMI B frequency. The installa- bremsstrahlung tagging system, located in the A2 area,
tion of the four HDSM magnets has been completed, and provides photons of known energy and ux. Circularly po-
commissioning is expected to start soon. larized photons can be obtained from the bremsstrahlung
As shown in table 1, important milestones parallel of polarized electrons, linearly polarized photons from an
to the technical developments were the establishment of oriented crystal radiator. The detection equipment is fo-
the Sonderforschungsbereiche ( Special Research Ini- cused on charged and neutral particle detection in a large
212 The European Physical Journal A
4 Selected experiments
The following sections will give examples for experiments
that are characteristic for the MAMI physics program and
Fig. 6. The Crystal Ball detector (left) and TAPS (right). that have had a large impact on the eld.
solid angle and energy range. The Saclay-built DAPHNE 4.1 Real Compton Scattering and the Polarizability of
detector (Detecteur a grande Acceptance pour la Physique the Proton
photo-Nucleaire Experimentale) uses a combination of
proportional wire chambers and layers of scintillation The electric and magnetic polarizabilities of the nucleon
counters and absorbers for charged particle and photon are static quantities that characterize the response of the
detection. A three-dimensional drawing is shown in g. 5. system to external electric and magnetic elds. Since the
The TAPS detector (original abbreviation for Two- highest elds that one can produce in the laboratory are
Arm Photon Spectrometer) can be arranged in dierent much too weak to have a measurable inuence, the best
congurations. Its 528 BaF2 crystals give good energy res- approach is to derive the polarizabilities from the energy
olution for photon detection. Charged particles can be and angular dependence of real photon scattering (RCS)
identied via the ratio of fast and slow scintillation light. at low photon energies. The experiments are challenging
The newest addition to the experimental arsenal is the since the cross sections are very small and, above pion pro-
Crystal Ball detector which has seen prior service at high- duction threshold, there is a large background of photons
energy facilities like SPEAR, DORIS, and the BNL AGS. from 0 decays.
Its central detector consists of 672 NaI crystals, again opti- The experimental results obtained with the TAPS de-
mized for photon detection. The electronic readout system tector and the bremsstrahlung tagging system [3] require
has been modernized and equipped with 80 MHz ash- signicant theoretical corrections and interpretation (see
ADCs. In the rst experiment, the Crystal Ball will be g. 8 for an example).
used in combination with TAPS as a forward detector (see The cross section is dominated by scattering o the
g. 6 for a picture of the setup) to measure the + mag- charge and the magnetic moment of the proton. In addi-
netic moment via the angular distribution of the decay tion, the incident and outgoing photons can couple to an
photons in the + + transition (from the high-mass exchanged pion. Finally, the polarizabilities enter linearly
tail of the + to its low-mass tail). only in the low-energy expansion. In practice, higher-order
Of particular importance for studying the spin degrees- terms need to be incorporated. A dispersion relation anal-
of-freedom of the nucleon has been the addition of the ysis [4] of the entire body of Compton scattering data
Bonn frozen-spin polarized H and D target (see g. 7 for shows that the proton is a very sti objects, i.e. it does
a picture). Its low magnetic holding eld and open geome- not deform much under the inuence of external static
try make it an ideal match for large acceptance detectors. elds.
B.A. Mecking: Twenty years of physics at MAMI What did it mean? 213
Dispersion relation
N contribution
asymptotic + N
asymptotic contribution
Fig. 8. Photon energy dependence of the Compton scatter- Fig. 9. Q2 -dependence of the generalized polarizabilities
ing cross section at a scattering angle of 135 . The dispersion E (Q2 ) and M (Q2 ).
relation analysis is the thin black line marked DR.
Fig. 11. LQCD results for the proton magnetic moment. The
data points give the results of the calculations for dierent val- Table 3. GDH Results from ELSA and MAMI.
ues of m 2 . The physical values of p and m 2 are marked by Source of Information k [MeV] IGDH [b]
the red cross. The short blue arrow indicates the typical PT
extrapolation range; the dotted blue arrow indicates the typical MAMI 200 800 226 5 12
LQCD extrapolation range. The best theoretical extrapolation PRL 87 (2001) 022003
is given by the green solid line. ELSA 800 2900 27.5 2.0 1.2
PRL 93 (2004) 032003
Low-energy extrapolation 140 200 27.5 3
calculations are performed far away from the physical pion MAID (2003)
mass. The generally accepted technique to extrapolate to High-energy extrapolation 2900 14
PLB 450 (1999) 439
the physical pion mass is to use the functional form for
the m -dependence given by PT. As an example, g. 11 Experimental sum 212 5 12
shows the result of a LQCD calculation [6] for the mag- Theoretical GDH integral 205
netic moment of the proton as a function of m 2 . With
the chiral extrapolation, the LQCD result gets close to the
true value.
experimentally) and above 2900 MeV were estimated the-
oretically and added to the experimental sum.
For the MAMI GDH experiment [7], circularly polar-
4.4 Experimental test of the GDH sum rule ized tagged photons from the bremsstrahlung of polarized
electrons were hitting a longitudinally polarized frozen-
The GDH sum rule for the nucleon was derived in 1966 by spin target. The DAPHNE detector in combination with
Gerasimov and, independently, by Drell and Hearn. The a forward scintillation counter and shower detector was
sum rule is based on fundamental assumptions: Lorentz used to measure the total cross section via identifying and
and gauge invariance, unitarity, and no-subtraction dis- adding up the individual reaction channels. In addition to
persion relations. The sum rule links the weighted integral testing the GDH sum rule, this experimental technique
over 3/2 (k)1/2 (k) to the anomalous magnetic moment also allows the measurement of the energy and angular
of the nucleon. 3/2 (k) and 1/2 (k) are the total hadronic dependence of the helicity-separated dierential cross sec-
photoproduction cross sections for total helicity 3/2 and tions for the individual reaction channels that contribute
1/2, respectively, as a function of the photon energy k. The to the total cross section.
integral over k extends from pion production threshold to From the MAMI and ELSA data, the GDH integral
innity, the weight factor is 1/k. has been calculated. It is shown as a function of the upper
This fundamental sum rule had never been tested due integration limit in g. 12. Table 3 gives the partial inte-
to the lack of appropriate experimental facilities. The test gral for dierent energy ranges; in the table, the statistical
of the sum rule requires a circularly polarized photon error is given rst, the second entry is the systematic error.
beam and a longitudinally polarized target, both with high Also listed are the contributions from photon energies be-
polarization. In addition, a detector to measure the total low and above the range that was covered experimentally.
cross section reliably needs to be available. Note that, because of the 1/k-weighting, the contribution
A major eort has been launched in Europe to test from threshold to 200 MeV is about the same as the con-
the validity of the GDH sum rule. For the proton, the low- tribution from 800 to 2900 MeV. Within the errors, the
energy part of the sum (200 800 MeV) has been measured nal experimental value agrees well with the theoretical
at MAMI, the contribution between 800 and 2900 MeV prediction, thus verifying the GDH sum rule.
has been investigated at the ELSA accelerator in Bonn. First preliminary data from a polarized deuteron tar-
Contributions below 200 MeV (that are di cult to access get are available. Extracting the GDH sum for the neutron
B.A. Mecking: Twenty years of physics at MAMI What did it mean? 215
Fig. 17. Setup to use quasi-monochromatic transition radiation for K-edge imaging. The sample containing the Mo foil is
located in front of the pn-CCD detector.
For higher N resonances, the extraction of the transi- applications, e.g. in the material sciences, in biology, and
tion form factor becomes quite involved: in the theoretical in medicine.
description, Born terms, unitarity, and channel coupling The most powerful X-ray sources in the energy regime
need to be taken into account. A full PWA is presently not of interest (K-edge of oxygen at 0.53 keV to the K-edge
possible due to lack of data; especially polarization data is of iodine at 33.16 keV) are dedicated synchrotron radia-
missing. To compensate for the lack of experimental data, tion facilities. Modern electron accelerators with their low
the analysis is often constrained by assuming the energy emittance electron beams may oer an attractive alter-
dependence of the resonance excitation. native. The interaction of these beams with matter have
A particularly interesting example is the N (1232) the potential of producing high brilliance X-rays with a
transition. A full partial wave analysis is possible since the tunable time structure. Processes of interest are transi-
(1232) is an isolated resonance, and the Watson theorem tion radiation, parametric X-rays, undulator radiation,
constrains the phases of the helicity amplitudes. The spin the Smith-Purcell eect, and channeling radiation.
1/2 3/2 transition is dominated by spin-ip M 1; how- An illustrative example is the use of the 855 MeV
ever, non-zero E2 and C2 multipoles are possible. This MAMI beam hitting a stack of 30 polyimide foils to pro-
would be a signature for a non-spherical charge distribu- duce hard X-rays via transition radiation [14]. Using a
tion in the (1232) or could be caused by the virtual highly oriented pyrolytic graphite crystal, an X-ray beam
photon coupling to a pion cloud. of about 20 keV was prepared; its two-dimensional spa-
At MAMI, a major experimental eort [3,12] was laun- tial distribution was measured in a pn-CCD detector (see
ched to determine the E2/M 1 ratio for the N (1232) g. 17 for details). Due to the crystal monochromator,
transition starting at the real photon point and extend- the X-ray beam had a correlation between position and
ing the measurements to virtual photons. For real pho- energy. By synchronously changing the electron beam di-
tons, the experiments used the tagging system to pro- rection and the crystal position, the X-ray energy spec-
duce linearly polarized photons and TAPS to detect the trum could be swept as a function of time. This technique
0 decay. For electroproduction, the three-spectro- avoids making the X-ray beam monochromatic with a slit
meter setup was used to measure the ep e p( 0 ) pro- system, and thus does not reduce the ux. Using the X-
cess. Tilting the proton spectrometer provided access to ray beam for K-edge imaging as a demonstration project,
out-of-plane observables. a 2.5 m Mo foil hidden in a 100 times thicker copper foil
The Q2 -dependence of the E2/M 1 ratio is shown in could be detected. Once perfected, this technique may be
g. 16. The results show that the E2/M 1 ratio is small used to image the human lung using xenon (mixed with
(around 2%) and negative. This nding is at variance with oxygen) as an absorber.
all models that consider constituent quarks, only. Models
that explicitly include the pion cloud can explain the data.
5 Experiment Theory interplay at MAMI
4.7 Production of low-energy radiation
As already pointed out in the discussion of the experimen-
In the context of the applied physics program [13], tech- tal program, there is a very close and eective collabora-
niques were developed that make use of the high-quality tion between the theorists working in the Institute and
MAMI electron beam for the production of high brilliance the experimentalists using MAMI. This tight coupling in-
X-rays. In addition to clarifying fundamental aspects creases the impact of the MAMI experimental program in
of radiation production, there are potential practical two important ways.
218 The European Physical Journal A
Fig. 18. Photos of the retirees. Top row from left to right: Hartmuth Arenhovel, Thomas Walcher, and Karl-Heinz Kaiser;
bottom row: Dieter Drechsel, Jorg Friedrich, and Hartmut Backe.
First, it provides the necessary corrections and facil- 6 MAMI funding and operation
itates the physics interpretation for quantities for which
the measurement strategy is clear and unambiguous. Typ- Securing steady funding for operating a major electromag-
ical examples are the calculations of the corrections to the netic nuclear physics facility within a university environ-
measured Gne values, or the corrections to the GDH in- ment has been a real challenge. Contributions from the
tegral to account for the unmeasured angular and energy University of Mainz, the state of Rhineland-Palatinate,
range. and from the Federal Government are required to keep the
operation of MAMI nancially viable. A large fraction of
Second, theoretical support starting in the early phase the federal funding is provided by the DFG in the frame-
of the experiment makes it possible to attack the deter- work of the SFB which is meant to support new ventures
mination of quantities for which due to the existence of for a limited period of time, but not meant to support
strong competing channels the measurement and anal- steady-state operation. The success of MAMI in this di -
ysis strategy is not obvious. Typical examples include the cult funding environment is testimony to the skills of the
extraction of the nucleon polarizability from p p, the people in charge of the Institute.
pion polarizability from p + n, and the magnetic The MAMI physics program has been comprehensive
moment of the + (1232) from p p 0 . and more characteristic for a national facility than for
a university-based accelerator. No physics problem that
The tight coupling between theorists and experimen- was worth attacking has been left out, even when it re-
talists at MAMI is unusual and unique. It clearly has been quired additions to the experimental equipment and im-
a major contributor to the success of MAMI. The ingre- provements to the accelerator (e.g.: the parity violation
dients necessary for the tight coupling are not easy to program).
identify, and cannot easily be transfered to other facilities. International collaborators have played an important
Very likely, the early history of the MAMI project and the role at MAMI, contributing both ideas and experimental
personal inclination of the people involved on both sides equipment. They obviously felt welcome at the Institute;
have played an important role. the natural hospitality and curiosity of the population of
B.A. Mecking: Twenty years of physics at MAMI What did it mean? 219
Mainz and the surrounding areas have likely had a bene- 8 Summary
cial eect, too.
MAMI has been an ideal training ground for students The development of the physics program and the experi-
who could experience all stages of an experiment, from mental facilities at the Mainz Microtron MAMI over the
planning to analyzing and publishing the results. The stu- last twenty years has been reviewed. MAMI and its user
dents were also trained in developing and using sophis- community have been working at the forefront of electro-
ticated experimental equipment, a good preparation for magnetic nuclear physics. Novel electron accelerators and
those who went on to pursue careers outside of nuclear experimental equipment were developed and have been
physics. In addition to educating students at MAMI, a used for ground-breaking research into the structure of
generation of young researchers has been trained who have nucleons and nuclei.
now gone out and successfully competed for faculty posi- The six people, Hartmuth Arenhovel, Hartmut Backe,
tions. Dieter Drechsel, Jorg Friedrich, Karl-Heinz Kaiser, and
Thomas Walcher (g. 18) who have retired (or are about
to retire) have played key roles in this development. They
7 The legacy of 20 year of MAMI physics have every right to be proud of what has been accom-
plished. I would like to take this opportunity to wish them
What this 20-year period at MAMI will be remembered Happy Retirement .
for will depend on the range of interests of the person
asking the question.
The development of the single-sided microtrons, References
MAMI A and B, has already changed the textbooks on
accelerators; the design and construction of the double- 1. A. Jankowiak, these proceedings.
sided MAMI C microtron will complete the microtron de- 2. A. Jankowiak et al., prepared for the 8th European Particle
velopment. Accelerator Conference (EPAC 2002), Paris, France, 3-7
The quality of the experimental data will not be sur- June 2002.
passed for a long time. Note that very often the accuracy 3. R. Beck, these proceedings.
of the nal answer is limited by the accuracy of the theo- 4. M. Vanderhaeghen, these proceedings.
retical corrections. Better accuracy can only be achieved 5. N. dHose, these proceedings.
by an improvement of both the experimental results and 6. R. Young (Jeerson Lab), private communication.
7. A. Thomas, these proceedings.
the theoretical interpretation.
8. H. Arenhovel et al., Z. Phys. A 331, 123 (1988).
Twenty years of MAMI physics have demonstrated
9. M. Ostrick, these proceedings.
that a tight coupling between theory and experiment can 10. F. Maas, these proceedings.
be mutually benecial. This coupling, which may not be 11. J. Friedrich, T. Walcher, Eur. Phys. J. A 17, 607 (2003),
easy to reproduce at other facilities, has been a major arXiv:hep-ph/0303054.
contributor to the success of MAMI. 12. H. Schmieden, these proceedings.
Finally, 20 years of MAMI physics have demonstrated 13. W. Lauth, these proceedings.
that it is possible although with a lot of eort to 14. F. Hagenbuck et al., IEEE Trans. Nucl. Sci. 48, 843 (2001).
operate a major facility within the framework of a German
university.