EPJ manuscript No.

(will be inserted by the editor)

G0 Experiment Status
G. Batigne for the G0 Collaboration
Laboratoire de Physique Subatomique et Cosmologie, 38028 Grenoble, France

Received: date / Revised version: date

Abstract. The G0 project is a parity violation experiment dedicated to the measurement of the proton
weak and axial form factors by means of elastic electron scattering. Combining these weak form factors
with the known electromagnetic ones makes possible the extraction of the contribution of strange quarks
to the charge and magnetization distributions in the nucleon. After introducing the physics case, this paper
describes the G0 apparatus and the measurements planned. An engineering run of this experiment took
place recently ; first results are presented.

PACS. PACS-key discribing text of that key PACS-key discribing text of that key

1 Introduction

els and to gain a better understanding of the structure of

hadrons.

The nucleon is naively viewed as a set of three valence

Some experimental results have already given indications
quarks (uud for the proton and udd for the neutron) but
that the strange quarks contribute to the properties of
its internal structure is far more complicated. In Quanthe nucleon. The -term value in N scattering suggests
tum Chromodynamics (QCD), the quarks are bound by
a contribution of strange quarks of about 130 M eV to the
the strong interaction mediated by the exchange of gluons
nucleon mass [1]. From eN Deep Inelastic Scattering data,
which can fluctuate into quark-antiquark pairs (uu, dd,
strange quarks could contribute to the nucleon spin at the
ss, ...). Thus the nucleon is a set of valence quarks sur-

level of few percent (hN |s 5 s|N i ' 10% [2]). But those

rounded by a sea of gluons and quark-antiquark pairs.

results are still questionable due to theoretical assumpAt low energy, the strong interaction cannot be treated
tions or unmeasured contributions.
by the theory of perturbative QCD which implies that the
The purpose of Parity Violating (PV) experiments in elecstructure of the nucleon cannot be easily described. Extron scattering is to extract the contribution of strange
periments are then required to constrain theoretical mod-

G. Batigne for the G0 Collaboration: G0 Experiment Status

quarks to the charge and magnetization distributions of

the nucleon. The formalism and the principle of this mea0

surement are first presented. Then the G experiment is

radius which is defined as follows :

dGE
2
hr i = 6
dQ2 Q2 =0

(3)

As photons couple only to quarks in the nucleon, the form

described and first results from its engineering run are
factors can be expressed as the sum of the contributions
given.

from the different quark flavors, Gq,N

E,M . The sea quarks
are produced from the fluctuations of gluons in quark-

2 Strangeness and Parity Violation

antiquark pairs. The probability of this process is pro-

The extended structure of nucleons has been extensively

studied by electron scattering. The virtual photon, exchanged during the reaction, probes the internal electromagnetic structure of the nucleons. The elastic cross section (eN ) depends on Dirac and Pauli form factors, F1

portional to 1/Mq2 where q is the flavor of the produced

quarks. At the energy of the current experiments, one can
neglect the contribution of the heaviest quarks (c, b and
t) and express the electromagnetic Sachs form factors as
follows :
(,N )

GE,M =

and F2 respectively, which describe the electromagnetic

eq Gq,N
E,M

(4)

q=u,d,s
(,p)

structure of nucleons. These form factors depend only on

The values of the electromagnetic form factors GE,M and

the four-momentum carried by the virtual photon (Q2 ).

GE,M have been or are presently measured for Q2 values

By varying the value of Q2 , one can tune the probed spa-

up to 1 (GeV /c)2 . These form factors provide four rela-

tial dimension. One can also use the Sachs form factors,

tions but there are twelve unknowns (Gq,N

E,M ). To reduce

defined as follows :

this number to six, the hypothesis of isospin symmetry is

GE = F1 F2

GM = F1 + F2

(1)

(,n)

d,n
d,p
u,n
u
d
made : Gu,p
E,M = GE,M = GE,M , GE,M = GE,M = GE,M
s,n
s
and Gs,p
E,M = GE,M = GE,M . The measurement of the

2
where = Q2 /(4MN
) with MN the mass of the nu-

contribution of the strange quarks (GsE,M ) requires then

cleon. In the Breit frame, GE and GM are respectively

two additional independent combinations of Gq,N

E,M to solve

the Fourier Transforms of the spatial distributions of the

the system.

charge and magnetization in the nucleon [3]. These form

Similarly to the electromagnetic interaction, one can also

factors are related to the static properties of the nucleon

define Sachs form factors for the weak interaction, GE,M

(at Q2 = 0) :
G,N
E (0) = QN

(Z,N )

(Z boson exchange in eN scattering). As the Z boson also

G,N
M (0) = N

(2)

couples only to quarks in the nucleon, the weak form factors of the proton can be written as :

where QN is the charge and N the magnetic moment of

the nucleon N . Another quantity of interest is the charge

(Z,p)

GE,M =

X
q=u,d,s

CVq GqE,M

(5)

G. Batigne for the G0 Collaboration: G0 Experiment Status

where CVq = 2T3 4Qq sin2 W is the weak vector charge

Because of the smallness of the weak interaction, the

of quarks of flavor q [5]. As the weak interaction violates

asymmetry is of the order of 105 but this time many nor-

parity symmetry, there is also an axial form factor GeA .

malization terms (i.e. luminosity, acceptance, ...) cancel in

The two weak form factors of the proton represent then

the ratio. This parity violation asymmetry is a function of

two new independent equations. The strange contributions

electromagnetic, weak and axial form factors [5,6] :

to charge and magnetization can then be extracted using

GF Q2 fE GE
AP V (Q2 , e ) =
4 2

(Z,N )

(Z,N )

+ fM GM
AD

+ fA GeA

the following formula [4] :

(8)
where
(,p)

(,n)

(Z,p)

GsE,M = (1 4 sin2 W )GE,M GE,M GE,M

(6)

Thus the determination of strange quark contributions re-

(,p)

AD

(,p)

(,p)

, fE = GM , fA = 0 GM

2
(,p)
(,p)
= GE
+ GM

fE = GE

(9)

The kinematical factors , and 0 are function of Q2 and

quires the measurement of weak form factors of the proton,

e , the electron scattering angle in the laboratory frame.
as, contrary to electromagnetic form factors, these weak
form factors are still mostly unknown.
The elastic scattering (eN ) cross section depends in fact

Formula 8 can also be expressed as a function of GsE , GsM

and GeA :

on both electromagnetic and weak form factors. But for

AP V = As=0 + GsE + GsM + GeA

values of Q2 around 1 (GeV/c)2 , the weak element ma-

where As=0 , the value of the asymmetry with no strange

trix, MZ , is 105 smaller than the one associated to the

quarks contribution, is derived from the Standard Model

exchange of a virtual photon, M . The experimental sys-

prediction and the measured electromagnetic form factors.

tematic errors related to normalization prevents then the

The terms , and depend on kinematics of the elec-

weak form factors to be obtained from unpolarized cross

tron scattering and electromagnetic form factors. Formula

section measurements. Contrary to the electromagnetic in-

10 shows that the asymmetry is sensitive to a linear com-

teraction, the parity symmetry is violated in the weak

bination of GsE , GsM and GeA . Thus for a given value of

interaction. Consequently the cross section of elastic scat-

Q2 , a set of three measurements must be performed to

tering of longitudinally polarized electrons on nucleons

separate them. The first two measurements correspond

(
e N ) depends on the helicity of the incident electrons

to electron scattering on proton in the forward (AF ) and

(i.e. + 6= , where refers to the two possible helic-

backward (AB ) directions. The axial form factor, GeA , can

ity state). Thus the weak part of the scattering can be

be calculated from theoretical predictions. However due

extracted by a parity violation asymmetry measurement :

to difficultes of estimating its value and Q2 dependence,

(10)

it appears worth also measuring it. Thus a third measureAP V

+
=
+ +

(7)

ment (AD ) is performed on deuterons at backward angles

G. Batigne for the G0 Collaboration: G0 Experiment Status

Table 1. Values in part per million (ppm) of As=0 , , and

tromagnetic form factors, etc.). The second experiment,

for the kinematical condition of the G0 experiment and for

HAPPEX at Jefferson Laboratory, measured the asym-

Q2 =0.5 (GeV/c)2 .

metry on LH2 at Q2 =0.48 (GeV/c)2 and forward scattering angle (e = 12.5 ) [9,10]. This experiment extracted a

As=0 (ppm)

(ppm)

(ppm)

(ppm)

AF

13

-16.7

60.8

25.0

1.4

AB

110

-29

18.0

40.3

8.5

AD

-39.8

14.7

9.0

10.1

110

linear combination of GsE and GsM :

GsE + 0.392GsM = 0.025 0.020 0.014

(13)

The effect of strange quarks is small but the individual

electric and magnetization contributions of strange quarks
cannot be separated in a single measurement ; also a lim-

in quasi-elastic scattering. The different values of As=0 ,

ited range of Q2 is presently covered. An experimental

, and are reported in table 1 for Q2 =0.5 (GeV/c)2

effort is thus underway. The PVA4 experiment [11] at

and for the electron scattering angles used in the G ex-

Mainz will separate GsE and GsM at Q2 =0.225 (GeV/c)2 .

periment. From the values quoted in this table, one can

Happex2 on the proton and an experiment on 4 He (E91-

see that the measurement at forward angle (AF ) is the

004), both at JLab, will be able to make the separation of

most sensitive to GsE . The backward angle measurement

GsE and GsM at the same Q2 as the SAMPLE experiment.

on protons (AB ) is more sensitive to GsM and a measure-

This contribution details the G0 experiment which is now

ment on deuterons (AD ) gives a way to extract the axial

starting to take data. The aim of this project is to sepa-

form factor.

rate GsE , GsM and GeA for three values of Q2 : 0.3, 0.5 and

Two parity violation experiments have already published

0.8 (GeV/c)2 [12]. The expected precision on GsE , GsM and

results. The first one, SAMPLE at MIT-Bates, measured

GeA and the wide range of Q2 covered by this experiment

backward angle scattering angle (130 < e < 170 )

can be seen to be quite unique in figure 1.

asymmetries on LH2 at Q = 0.1 (GeV/c) [7]. A second

Theoretical models have been used to predict the contri-

measurement was performed at the same kinematics but

butions of strange quarks [13]. Models are based on chiral

on deuterium [8]. The combination of the two results gives

perturbation theory [14], Vector Dominance Meson ap-

the following values for GsM and the isovector part of GeA :

proches [15] and lattice QCD techniques [16] and provide

GsM = 0.14 0.29 0.31

(11)

GeA (T = 1) = 0.22 0.45 0.39

(12)

values of GsE and GsM . Some of these predictions are reported in figure 1, compared with projected experimental results. One can see that their models give different

where the first error corresponds to statistics and the

strange quark contributions. So experimental results are

second to systematics (errors on beam properties, elec-

required to constrain these theoretical models. About the

G. Batigne for the G0 Collaboration: G0 Experiment Status

isovector axial form factor, its value at Q2 = 0 (GeV/c)2

0.6

was calculated by Zhu et al [17] and its Q2 dependence

0.4

PT
Lattice QCD
3 poles
4 poles
6 poles

0.2
GEs

by Maekawa et al [18]. Their results were combined and

are reported in figure 1. The only published experimental

-0.2

result is given by the SAMPLE collaboration and has a

-0.4

G proposed
4

He proposed

large error. In these figures are also reported expected er-

-0.6

rors for 4 He, Happex2 and G0 experiments. So this set of

0.6
0.4

els.

0.2
GMs

new experiments will strongly constrain theoretical mod-

0.2

0.4 0.6 0.8

2
2
Q (GeV/c)

1.2

PT
Lattice QCD
3 poles
4 poles
6 poles
(HAPPEX II+4He) proposed
0
G proposed
SAMPLE 2000

-0.2

3 G Experiment

-0.4
-0.6

lation experiments are of the order of 105 . However one

requires an absolute precision of 106 107 , and thus

GAe(T=1)

The values of asymmetries to be measured by parity vio-

0.2

0.4 0.6 0.8

2
2
Q (GeV/c)

1.2

SAMPLE 2000 [7]

G0 Proposed

all false asymmetry sources must be controlled to high

0

precision. There are thus general requirements that must

be fulfilled for a violation parity experiment. This is in
particular the case for the beam properties. First the po-

-1
Zhu et al. + dipole
Zhu + Maekawa (s=-1 =0.55)
Zhu + Maekawa (s=+1 =0.55)

-2

0.2

0.4

0.6

larization has to be known at the few percent level because

0.8
1
Q2 (GeV/c)2

the measured asymmetry is proportional to the beam polarization. The helicity reversal must happen with a high
frequency (30-60 Hz typically) to minimize the effects of

Fig. 1. Predictions of some theoretical models on GsE , GsM

and GeA (T = 1). The expected errors of approved experiment
at JLab and the results of SAMPLE are also shown (figures

slow drift, especially on intensity. Feedback systems on

taken from [19], references therein).

intensity, energy and position are used to keep the beam

properties as close as possible in the two helicity states
full data acquisition event by event is therefore excluded.

and thus to limit corrections to be applied in the asymmetry calculation. To achieve a statistical precision of 107 ,
one has to record 10

14

events, requiring a large luminos-

ity. The parity violation experiments use long (between 20

and 40 cm) cryogenic targets and high beam intensity. A

The measurement is made by the integration of a signal or

summing events in histograms. Also fast detectors (around
10 ns) are used. The dead time of electronics and acquisition must be as low as possible.

G. Batigne for the G0 Collaboration: G0 Experiment Status

early 2004. It will measure the parity violation asymmetry

for Q2 values ranging between 0.1 and 1 (GeV/c)2 using
a single beam energy (3 GeV ). The shape and position
of each detector are designed in such a way that it covers a given range in Q2 . The scattering angle of electrons
is then varying between 6 and 22 . As the detection of
electrons is very difficult at these angles, the recoil protons
0

Fig. 2. The G apparatus.

instead are detected (48 < p < 78 ). The background

formed by neutrals ( and neutrons) is mostly removed by

The G0 experiment that is being described in this con-

a set of internal collimators preventing direct view of the

tribution is a parity violation experiment which is taking

target. The background formed by charged particles ( +

place at Jefferson Laboratory (JLab, Virginia, USA). The

and inelastically scattered protons) is rejected by a Time

collaboration is formed of North-American, Armenian and

of Flight (ToF) measurement. For each FPD detector, a

French institutes and universities [12]. This experiment is

ToF spectrum is constructed. Figure 3 shows an example

installed in Hall C of JLab. The CEBAF accelerator de-

of a simulated ToF spectrum. The beam time structure is

livers beam with 75% polarization and 40 A intensity.

chosen so that all particles produced by a beam burst ev-

This beam is pulsed with a period of 32 ns. The G0 de-

ery 32 ns are detected before the arrival of the next beam

tector is first composed of a 20 cm long target which can

burst. The elastically scattered protons can be selected by

be filled with liquid hydrogen or deuterium. This target is

cuts on ToF. Custom electronics were designed and allow

located at the center of a toroidal superconducting mag-

to discriminate photomultiplier signals, make the coinci-

net. The produced field bends charged particles on a focal

dence between the two layers of scintillator for each de-

plane where are located a set of detectors (FPD1 ) formed

tector2 and generate the ToF spectra. These histograms

of scintillator pairs, optically coupled to phototubes. Eight

are read during each helicity reversal which occurs at a

sectors of detection (octants) are used, each formed of 16

frequency of 30 Hz. This procedure eliminates deadtime

FPD detectors. The set-up is shown in figure 2. In order

related to the data acquisition.

to separate the weak form factors, the G0 experiment is

The second phase of the G0 experimental program will

scheduled into two distinct set of measurements. The first

be dedicated to back angle measurements (both on LH2

phase is dedicated to the measurement at forward scatter-

and LD2 ) and should take place in 2005-2006. In contrast

ing angles (AF ) and the final data taking will take place

with the first phase, electrons are detected. Their scat-

Focal Plan Detector

This allows additional rejection of neutrals.

G. Batigne for the G0 Collaboration: G0 Experiment Status

CED

LH2 - backward - E0 = 539 MeV

9
8
7

10 3

6
5
4

Elastic electrons
Inelastic
electrons

10 2

2
1
0

10

15

20

25

30

10

12

14

16
FPD

Fig. 3. Example of a simulated ToF spectrum. The vertical

Fig. 4. Simulated coincidence matrix between CED and FPD

axis represents relative counting rates and the horizontal axis

detectors.

ToF is expressed in ns.

the counting rates corresponding to the matrix locations

tering angle is around 110 and the value of Q2 varies

associated to elastically scattered electrons.

slowly in the acceptance of the collimators. Therefore the

Pions are however also detected in some of the matrix

measurements will be performed at three beam energies

locations associated to elastically scattered electrons. To

(420, 580 and 800 MeV) and values of Q2 equal to 0.3,

discriminate and reject these pions, which rates would

0.5 and 0.8 (GeV/c)2 will be accessible. To detect parti-

become very high with a deuterium target, an aerogel

cles scattered at backward angles, the overall apparatus

Cerenkov
detector will be installed close to the CED ho-

must be turned around. The detected electrons are ultra-

doscope (see figure 5). The refractive index of the aerogel

relativistic so a ToF measurement cannot be used to re-

is chosen so that only electrons produce light when pass-

ject the background3 . Thanks to the magnetic field of the

ing through the Cerenkov

detector. The Cerenkov
signal

spectrometer, elastically scattered electrons have different

is consequently used to validate the encoding of an event.

trajectories compared to inelastically scattered electrons

This procedure allows the rejection of typically 95% of

and . So eight hodoscopes of nine additional scintilla-

the pions. The Cerenkov

detectors are under construction

tor detectors, called CEDs4 , will be installed at the exit

and test. New custom electronics is presently designed and

of the magnet cryostat and associated to each octant of

will form the coincidence matrix and process the Cerenkov

FPD. A coincidence matrix between CED and FPD detec-

signal.

tors allows one to separate elastic and inelastic events as

shown in figure 4. The asymmetry will be calculated from
3

In this part of the experiment the charged background is

and inelastically scattered electrons.

4

4 Engineering run

Cryostat Exit Detectors

The first engineering run of the G0 experiment took place

from October 2002 to January 2003 in the forward angle

G. Batigne for the G0 Collaboration: G0 Experiment Status

boiling. A key point for the experiment concerned the opCerenkov

eration of the spectrometer. The nominal value of current

was achieved in its superconducting mode. The magnetic

FPD

field is directly related to the Q2 range measured by the

CED

experiment (0.1 < Q2 < 1 (GeV/c)2 ). The first part of

Collimator

Incident Electrons

target

this engineering revealed that the hall background was

higher than expected. It was found that the main source
Fig. 5. Sketch of one octant of detectors in the back angle
configuration. The collimators are used to select the kinematical range but also to shield detectors from the line of sight of
the target.

of background were neutrals generated at the beam pipe.

So the shielding was properly improved. Concerning the
detectors, the high voltages for photomultipliers (PMT)
were set at values allowing high detection efficiency and

configuration (AF ). During this period, not only the apparatus was tested but also many special features required
of the CEBAF beam. First the accelerator succeeded to
deliver the G0 beam with use of a new laser system. The
standard CEBAF beam is pulsed with a 2 ns period with
an intensity up to 100 A. The challenge was to build a
40 A beam with a 32 ns period which means to have 6
times more electrons than usual per burst. Various feedback systems were also tested. The goal is to achieve, for
the 700 h of data taking scheduled for the production run,
an overall charge asymmetry less than I/I < 1 ppm

the PMTs were able to stand rates at nominal beam current. The gains were matched and their stability was satisfactory over time and different beam conditions. Discriminator thresholds were adjusted in order to remove the
noise and low energy particle background while keeping
100% efficiency for elastic proton detection. All electronic
channels worked fine and ToF spectra were found stable
in time. An example of ToF spectrum is given in figure
6 (the vertical lines in this figure represent typical ToF
cuts to select elastic protons). Some studies are still under progress, in particular on the deadtime corrections.

and differences in position less than X < 20 nm. The

amount of useful data for asymmetry calculation recorded

The results of this engineering run also show that

during this engineering run represents only 1/16 of the fi-

the rates of inelastic events were higher than expected

nal statistics. For these data, the charge asymmetry was

from the simulations. The main source of these addi-

less than 5 ppm and the position asymmetries less than

tional events was assigned to reactions occuring in the alu-

50 nm which meet the requirements for these recorded

minium target cell windows. So the thickness of these win-

data. The target system was also tested ; in particular

dows will be reduced for the next engineering run sched-

we checked that it can stand the beam intensity without

uled this fall.

G. Batigne for the G0 Collaboration: G0 Experiment Status

15
14

Tof difference (ns)

13
12
11
10
Octant 2 - shift of octant

Octant 4 - shift of octant

Octant 6 - shift of octant

Octant 8 - shift of octant

Fig. 6. Example of a measured ToF spectrum. The time bin-

6
0

Simulation @ 5000 A

8
10
Detector

12

14

16

ning is 0.250 ns.

Fig. 7. Comparison of the ToF difference between elastic pro-

A check of agreement between simulation, and the under-

ton and + extracted from the simulation (black line) and

standing of the set-up, and measurement has been made

measurement (points).

using the ToF difference between pions and elastic protons. Figure 7 shows the results of this study for each
detector. The observed agreement is quite good (around
100 ps) and required the inclusion of detailed geometry
and event generators in the simulation. This ToF difference is also important as it allows to determine the mean
2

value of Q associated with each FPD detector. Indeed

pions are ultrarelativistic with almost fixed ToF contrary
to protons for which the kinetic energy is proportional to
Q2 . In order to keep the systematic error introduced by
2

the uncertainty on Q below 5% of the statistical error,

the value of Q2 must be known with a precision of 1%.
This is achieved with the observed precision of 100 ps on
T oF .
The parity violation asymmetries have been calculated for

If there is no false asymmetry generated from the beam

or the electronics the results with and without this IHWP
must be exactly the same but with opposite sign. Figure
8 shows that it is the case within errors (only statistical
ones are displayed). The analysis is still underway. The
asymmetry of inelastic events has to be subtracted from
the measured asymmetry in order to extract the elastic
proton asymmetry. Another study performed at present is
related to the electronics response to the charge asymmetry. The goal is to achieve a systematic error at the order
of 1% of the charge asymmetry. The main contribution to
this systematic error is the deadtime. The understanding
of the beam properties and feedback systems is also in
progress. Improvements and changes will be tested during
the next engineering run this fall.

each detector. The statistical error associated with this

engineering run is about four times larger than the one

5 Conclusions

expected in production run. A Half Wave Plate (IHWP)

can be inserted at the injector to reverse the electron helic-

The purpose of parity violation measurements in (eN )

ity compared to the helicity signal sent to the electronics.

elastic scattering is to determine the contribution of

G. Batigne for the G0 Collaboration: G0 Experiment Status

10

Asymmetry (ppm)

5. M. Musolf et al., Phys. Rep. 239, (1994) 1

6. P.A. Souder and K.S. Kumar, Prog. Part. Nucl. Phys. 45,
(2000) S333
7. R. Hasty et al., Science 290, (2000) 2117
-5

8. D. Spayde et al., Phys. Rev. Lett. 84, (2000) 1106

-10
-15

9. K.A. Aniol et al., Phys. Rev. Lett. 82, (1999) 1096

2

10

12

Detector

14

Fig. 8. Comparison of the measured asymmetries with and

10. K.A. Aniol et al., Phys. Lett. B 509, (2001) 211

11. F. Maas et al., proceeding of Parity Violation In Atoms
and Polarized Electron Scattering, ed. B. Frois and M.A.

without the Insertable Half Wave Plate (IHWP).

Bouchiat, (1999)

strange quarks to the charge and magnetization distributions of the nucleon. Several experimental programs
are dedicated to these measurements : SAMPLE at MIT-

12. G0 web page : [Link]

13. D.H. Beck and B.R. Holstein, Int. Jour. Mod. Phys. E 10,
(2001) 1

Bates, Happex at JLab, PVA4 at Mainz and G at JLab.

14. T.R. Hemmert, B. Kubis and U.-G. Meiner, Phys. Rev. C

The special contribution of the G experiment is that it

will be the first experiment to extract

GsE ,

GsM

and

GeA
2

over a large range of Q ( = 0.3, 0.5 and 0.8 (Gev/c) ).

The results from the first engineering run demonstrate a

60, (1999) 045501

15. H. Forkel, Phys. Rev. C 56, (1996) 510
16. S.J. Dong, K.F. Liu and A.G. Williams, Phys. Rev. D 58,
(1998) 074504

good understanding of the beam properties, detector and

17. S. Zhu et al., Phys. Rev. D 62, (2000) 033008

electronics responses. The physics analysis is under way

18. C.M Maekawa, J.S. Veiga and U. Van Kolck et al., Phys.

and the data taking production run will take place at the
beginning of the next year.

References
1. J. Glaser, H. Leutwyler and M. Saino, Phys. Lett. B253,
(1991) 252
2. K. Abe et al, Phys. Rev. D58, (1998) 112003
3. F. J. Ernst, R. G. Sachs and K. C. Wali, Phys. Rev. 119
(1960) 1105
4. D.H. Beck and R.D. McKeown, Ann. Rev. Nucl. 51 (2001)
189

Lett. B 488, (2000) 167 Q2

19. R. Tieulent, PhD Thesis, Grenoble University, ISN 02-27