Particle Physics: Ideas and Recent Developments

Particle Physics: Ideas and Recent Developments
NATO Science Series

A Series presenting the results of activities sponsored by the NATO Science
Committee. The Series is published by lOS Press and Kluwer Academic
Publishers, in conjunction with the NATO Scientific Affairs Division.
A. Life Sciences lOS Press

B. Physics Kluwer Academic Publishers
C. Mathematical and Physical Sciences Kluwer Academic Publishers
D. Behavioural and Social Sciences Kluwer Academic Publishers
E. Applied Sciences Kluwer Academic Publishers
F. Computer and Systems Sciences lOS Press
1. Disarmament Technologies Kluwer Academic Publishers

2. Environmental Security Kluwer Academic Publishers
3. High Technology Kluwer Academic Publishers
4. Science and Technology Policy lOS Press
5. Computer Networking lOS Press
NATO-PCO-DATA BASE
The NATO Science Series continues the series of books published formerly in the NATO ASI
Series. An electronic index to the NATO ASI Series provides full bibliographical references (with
keywords and/or abstracts) to more than 50000 contributions from international scientists
published in all sections of the NATO ASI Series.
Access to the NATO-PCO-DATA BASE is possible via CD· ROM "NATO-PCO-DATA BASE" with
user-friendly retrieval software in English, French and German (WTV GmbH and DATAWARE
Technologies Inc. 1989).
The CD-ROM of the NATO ASI Series can be ordered from: PCO, Overijse, Belgium
Series C: Mathematical and Physical Sciences - Vol. 555

Particle Physics:
Ideas and Recent Developments
edited by
J.-J. Aubert
CPPM, Universite de la Mediterranee
and IN2P3, CNRS,
Marseille, France
R. Gastmans
Katholieke Universiteit Leuven,
Leuven, Belgium
and
J.-M. Gerard
Universite Catholique de Louvain,
Louvain-Ia-Neuve, Belgium
"
111...
Springer-Science+Business Media, BV.

Proceedings of the NATO Advanced Study Institute on
Particle Physics: Ideas and Recent Oevelopments
Cargese, France
July 26-August 7,1999
A C.I.P. Catalogue record for this book is available from the Library of Congress.
ISBN 978-0-7923-6436-8 ISBN 978-94-011-4128-4 (eBook)

DOI 10.1007/978-94-011-4128-4
Printed an acid-free paper
AII Rights Reserved

© 2000 Springer Science+Business Media Oordrecht
Originally published by Kluwer Academic Publishers in 2000
Softcover reprint of the hardcover 1st edition 2000
No part of the material protected by this copyright notice may be reproduced or utilized in
any form or by any means, electronic or mechanical, including photocopying, recording
ar by any information storage and retrieval system, without written permission from the
copyright owner.
TABLE OF CONTENTS
Preface ................... . . . ............................... . ........ VB
Results from LEP 200 ............................... 0 . 000... 0 0. ...

0 0.. 1
E. Fernandez
Anomalous Quartic Couplings at LEP2 .. 0 00.. 00 . 0 ...
0 0.. 0.... 0 000.... 55
W.J o Stirling and A. Werthenbach
Physics with an e+ e- Linear Collider at High Luminosity .... 0000. 065
0. 0
P.M. Zerwas
Point Splitting and U(l) Gauge Invariance .... 00 ... 0.. 00000 ... 00 .. 000095
Do Olivie
Neutrino Oscillations 00. 0 0 0 0 0000. .......
0 0000... 00... 0. 00. 0 .. 000 ...
0 0101
Lo Di LelIa
Physics of CP Violation and Rare Decays o. 0... 0 . .....
0 0 0. 00.. 0 0 0 . 0171
00
T. Nakada
HERA - Experimental Results .. 0 0 000. .. ... 0.... 00 . 0. 00 ......
o 000 .. 0203
R.J 0 Cashmore
Recent Results and Perspectives at CDF and D00 0..... 00.0.00.00 ... . 213
HoE. Montgomery
Heavy Gauge Boson Production at Small Transverse Momentum in
Hadron-Hadron Collisions ........ 0 . 0.. 0... 0...... 0. ..........
0 00 0....
0 235
Ao Kulesza and WoJ. Stirling
Recent Developments in the Theory of Strong Interactions: Power
Corrections and Renormalons 0 0.. 0 . 0.. 0. 0. .....
0 0.. .. ........ 0 . 0.. 00 0 0245
0
GoP. Korchemsky
Supersymmetric Extensions of the Standard Model 0.... . .. .. 0 . 0 0..00267
F. Zwirner
Enhanced Global Symmetries and a Strong Electroweak Sector: o. 0.. 0329
ChirallConformal Phase Transition
F. Sannino
Supersymmetry and Duality in Field Theory and String Theory 0 0.. 0339
E. Kiritsis
VI
Baryogenesis ........................................................ 397

M. Shaposhnikov
Cosmology and Dark Matter ......................................... 417

P. Salati
Index ... .. ........... ... ................................. . .. ... ..... . 511
PREFACE
The 1999 Cargese Summer Institute on "Particle Physics: Ideas and Re-
cent Developments" was organized by the Universite de la Mediterranee,
Marseille (J.-J. Aubert), the Universite de Paris-Sud, Orsay (P. Binetruy),
CERN, Geneva (D. Froidevaux), the Katholieke Universiteit Leuven
(R. Gastmans), the Universite Catholique de Louvain (J.-M. Gerard), the
Ecole Normale Superieure, Paris (J. Iliopoulos), and the Universite Pierre
et Marie Curie, Paris (M. Levy). This school is the thirteenth Summer In-
stitute on High Energy Physics jointly organized at Cargese by three of
these universities since 1975.
For more than 25 years, the Standard Model in particle physics has
repeatedly been confronted with numerous experimental results. These ex-
periments reach higher and higher precision, yet the Standard Model with-
stands those tests with great success. This does not imply that the Standard
Model gives the answers to all questions about the ultimate constituents of
matter.
The program of this Summer School paid attention to these two as-
pects of the Standard Model: new experimental confrontations and recent
developments in its theoretical understanding.
The new experimental data originated not only from the big accelerator
experiments , such as those performed at CERN with LEP2, at DESY with
HERA and at Fermilab with the TEVATRON, but also from the under-
ground neutrino experiments, such as SuperKamiokande. Since the neutrino
experiments suggest the existence of neutrino oscillations, a critical analy-
sis of this result was imperative because of its farreaching consequences for
particle physics, the solar model and astrophysics in general. Another open
question, which was thoroughly discussed, is the origin of the observed CP-
violation and its possible description within the framework of the Standard
Model.
Cross-fertilization of particle physics, astrophysics and cosmology has
become increasingly frequent in the last decade or so. A discussion of the lat-
est insights on topics, such as baryogenesis, dark matter, neutrino masses,
large scale structures and the cosmological constant, was therefore in order.
This overview of the present state of the art in particle physics was
completed by a discussion of likely future developments in the construction
of new accelerators and in the critical examination of new ideas. Along these
lines, we mention the progress in non-perturbative QCD and the possibility
that supersymmetry leads us beyond the Standard Model.
Vll
viii
We owe many thanks to all those who made this Summer Institute
possible!
Special thanks are due to the Scientific Committee of NATO and its
President for a generous grant. We are also very grateful for the financial
contributions given by the C.N.R.S., by the Institut National de Physique
Nucleaire et de Physique des Particules (IN2P3), and by the Ministere de
l'Education Nationale, de la Recherche et de la Technologie.
We also want to thank Ms. D. Escalier for her efficient organizational
assistance, Ms. C. Ariano and Ms. B. Cassegrain for their smooth secretarial
efforts, Mr. D. Olivie for his valuable aid in preparing these proceedings,
Mr. J.-A. Ariano for his help in all material matters of the school, and, last
but not least, the people from Cargese for their hospitality.
Mostly, however, we would like to thank all the lecturers and partici-
pants: their commitment to the school was the real basis for its success.
Jean-Jacques Aubert
CPPM, Universite de la Mediterranee and IN2P3, CNRS
Marseille, France
Pierre Binetruy Daniel Froidevaux

Universite de Paris-Sud CERN
Orsay, France Geneva, Switzerland
Raymond Gastmans Jean-Marc Gerard

Katholieke Universiteit Leuven Universite Catholique de Louvain
Leuven, Belgium Louvain-Ia-Neuve, Belgium
Jean Iliopoulos Maurice Levy

Ecole Normale Superieure Universite Pierre et Marie Curie
Paris, France Paris, France
RESULTS FROM LEP 200
E. FERNANDEZ
Universitat Autonoma de Barcelona/IFAE
Campus Bellaterra, Edifici C
E-OB193 Beliaterra, Barcelona, Spain
Abstract. Results from the LEP experiments taking place at the LEP-
2 accelerator are presented. They include measurements of fermion pair
production above the Z peak, of WW and ZZ production cross-sections
and of the W mass, and searches for the Standard Model Higgs boson. The
significance of the measurements of the W mass in the context of precision
tests of the Standard Model is discussed .
1. Introduction
These lectures are about some of the results obtained with the LEP ac-
celerator in its second phase, called LEP-2 or LEP-200. The 200 refers to
the total center of mass energy, in GeV, of the e+e- collisions, an energy
only reached a few days ago (July 1999). There are four major experiments
taking place at LEP: ALEPH, DELPHI, L3 and OPAL, each of them be-
ing carried out by large collaborations of 400 people or more, from many
institutions and many countries. The amount of work done by these collab-
orations is enormous and has resulted in many publications, close to 150 per
experiment, so far. This reflects both the richness of the physics that can
be studied at LEP and the dedication of the many people involved. I do not
have the exact figure, but guess that the number of doctoral thesis written
about LEP is more than 1000 (I have the number for my own group, and it
is 15). Some of you in this audience know first hand. For this reasons it is
almost impossible to cover in three lectures all the results obtained at LEP,
unless one goes over them superficially. Here I have taken the approach
of selecting a few topics and try to explain not only the results and their
meaning, but also give a brief account of how they were obtained.
J.-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 1-54.
© 2000 Kluwer Academic Publishers.
2 E.FERNANDEZ
Lecture 1 is devoted to some of the topics that can be studied from

fermion-antifermion production, which is still the process with the highest
cross-section at LEP 2. Lecture 2 is about the measurement of the W mass
and its implications for electroweak precision results. Lecture 3 covers the
searches for "new" phenomena, namely the Higgs boson of the Standard
Model and supersymmetric particle searches. Lectures 1, 2 and the searches
for the Standard Model Higgs are summarized in this report while for su-
persymmetric particle searches I refer the reader to some of the excellent
reviews of this topic. But first, a few comments on the LEP accelerator.
2. The LEP accelerator
As for any other modern high-energy accelerator, the building of LEP took
many years. Early conceptual studies date from 1976 and resulted in a
machine able to reach a center of mass energy of 200 GeV, with a luminosity
of 1032 cm- 2 s- 1 and a circumference of 50 Km. Another early design,
more similar to the present LEP, consisted on a collider ring of 22 Km
circumference, with an energy of 140 GeV in a first phase, to be upgraded to
200 Ge V in a second phase. Several workshops were organized by CERN and
ECFA (European Committee for Future Accelerators) to study the physics
of such an accelerator, namely in Les Houches (1978), Rome (1978) and
Hamburg (1979). A conclusion from these studies was that the maximum
energy should be above the threshold for WW production. The LEP project
was finally approved by the CERN Council in 1981, and consisted in two
phases, with energies of 140 GeV and 200 GeV. It was the discovery of the
Wand Z particles at CERN in 1983, and the first measurements of their
masses, what finally fixed the parameters of the machine we have today.
The first phase, LEP-1, started with the first data taking on the 14th
of July of 1989, and was completed in the fall of 1995. The energy was kept
close to the Z mass during these years. The LEP-2 phase started in 1995,
with short runs at energies of 130-136 Ge V in the center of mass (the so-
called LEP 1.5 period), and reached the WW threshold in 1996. Since then
the machine has been running steadily and reached the 200 Ge V center of
mass energy very recently.
It should be emphasized that raising the energy of LEP from the 90 Ge V
of LEP-1 to the 200 GeV of LEP-2 is by no means trivial. The energy loss
by synchrotron radiation grows as ,41 R2, where, is the Lorentz gamma
factor of the electrons and positrons, and R is the radius of curvature of their
path. Therefore, at fixed radius and current, the energy loss grows as the
4th power of the energy. At 100 GeV an electron looses 3 GeV per turn of
LEP, which, for a gradient of 6 MV 1m requires 500m worth of accelerating
cavities. Assuming that the cavities have an efficiency of 10% this translates
RESULTS FROM LEP 200 3
in 240 MW of electric "plug" power. Furthermore the inefficiency of normal

copper cavities grows strongly as a function of the energy, which in practice
means that LEP-2 can only be built with superconducting cavities. The
energy of LEP-2 has been growing steadily every year (Table 1), limited by
the availability of superconductive cavities and the corresponding cryogenic
power.
Period Energy (Ge V) Luminosity (pb -1)
1995 130/136 6.2

1996 161 12.1
1996 172 11.3
1997 183 63.8
1998 189 196.4
1999 192 27 (up to July)
TABLE 1. Energies and integrated luminosities

of LEP-2, up to July 1999.
3. Fermion pair-production at LEP-2

The e+ e- annihilation cross section is described in first order by s-channel
exchange of a, and a Z boson, depicted by the Feynman diagrams of Fig. 1.
f
"Z
e f
Figure 1. First order Feynman diagrams for e+e- ~ If.
The total annihilation cross-section (see ref. [1]) e+e- -+ f f has a

huge peak at the Z resonance as seen in Fig. 2 for the specific channel
e+e- -+ /.l+f..L-.
The cross-section for the production of qij pairs reaches about 41.5 nb
(after corrections for initial state radiation) right at the Z peak. Well above
4 E. FERNANDEZ
<1~(nb) ALEPH
,--------------------------,
• I.EP (ALEPH)
o PEP
A PJ....RA
o TRISTAN
i
w1 t
t_
~
L........L ........L J
~
L _-'- __ , .--1--,---,--.-..1_.1.__.1..._
~ "
'--.~
~
__ l_.L . ........
m
J..... __ -'-.__ -,--_J .... ..L ....
~
"5' (G.V)
Figure 2. e + e - annihilation cross-section as a function of the collision energy, for the

specific channel e + e - -+ 11+11- .
this energy the cross-section is also dominated by fermion-pair production

(with sharp thresholds for WW and ZZ production), but typical cross-
sections are a factor of a thousand smaller than at the Z peak, thus resulting
in much smaller event samples than at LEP-l. This is shown in Fig. 3 [2]. To
fix ideas: well above threshold the WW production cross-section is 16 pb,
that is, every inverse pb of accumulated luminosity gives 16 WW pairs. We
are very far here from the huge data samples obtained at LEP-l. Table 1
also gives the accumulated luminosities obtained at the different energies at
which LEP-2 has operated. Fermion-pair production has been extensively
studied by the four LEP collaborations (see references [3] to [14]).
Initial state radiation has a dramatic effect at LEP-2. If after radiation
the remaining center of mass energy of the e+ e- pair is near the Z peak,
the cross-section is highly enhanced (see below). Such events are called
"radiative return events" . Ideally we would like to separate them from those
with no radiation ("non-radiative"). For the latter the, and Z exchange
diagrams are of comparable importance.
The procedure to select radiative events involves finding photon(s) and
jets of particles and, through kinematic fitting, test the consistency of initial
state radiation. Once the photon(s) is removed, the center of mass energy
after radiation, called 8', can be computed from the four-momenta of the
remaining particles. Thus two classes of events are defined:
(1) 8'/8 large (non-radiative)
-:D
~
b 10' ••••• a(e+e- ~ X) (pb)
··········•······ ........... Lqq (ISR)

Lqq ................... .
- ........
... _w .................................. ..
ww,~~~~------J
-,
10
120 140 160 240

v'S (GeV)
Figure 3. The e+ e- annihilation cross-section at LEP-2 energies
(2) s'/ s > 0.01 (inclusive)

where "large" , in most analysis, means greater than 0.85. The s' distribution
for e+e- -+ II at 183 GeV is shown in Fig. 4 for the OPAL experiment
[13].
Two usual complications with this procedure are the following:
(a) Initial and final state radiation cannot be separated and, further-
more, they interfere. This makes difficult the definition of S'. Fortunately,
the effect is small ( 1%) and can be "effectively removed" (calculated).
(b) Final states involving 4 fermions (J II}'), cannot be naturally sep-
arated from 2 fermion processes in some cases. This effect is taken into
account when comparing the data with the theory.
All the LEP collaborations have studied these type of events and mea-
sured cross-sections and forward-backward charge asymmetries for qij and
lepton-antilepton pair production. (See Figures 5 as an example of the mea-
surement of the cross-section and leptonic forward-backward asymmetries).
The results of the four LEP experiments are in very good agreement
with the Standard Model predictions. Comparison of the latest data, from
the four LEP experiments combined, with the SM expectations has just
6 E. FERNANDEZ
OPAL 183 GeV
10
50 100 150 200 100 150 200

.Js'/GeV s'IGeV
.l!l 160 80
"~
w
140 (c) ~+~.
.l!l
w
"'"> 70 (d) ,,+,,'
120 60
100 50
80 40
60 30
40 20
20 10
0 50 0 50
100 150 200 150 200
.Js'/GeV .Js'IGeV
Figure 4. s-prime distribution for e+e- -+ If from OPAL [13] .
ALEPH
- his)'n > 0. 1 trIA rn.or1o.f'r)
••. (S,!s)1/1 > 0_9 (Op.en lo4ort..,)
10
0, 6
0 ,6
0,'
O.:?
-, 0
'0 - 0 .2
- 0 ,'
-2 -0.6
10
-0.8
-I
60 80 100 120 140 '60 180 60 80 100 I~O 140
E,~(GeV)
Figure 5. Measurements of the cross-sections for inclusive and non-radiative events and
of the forward-backward leptonic (for J1, and 'T) asymmetries [5].
been presented last week at the European Physical Society HEP Conference
in Tamp ere , Finland (ref. [15]), and are shown in Fig. 6.
Another comparison with the Standard Model was done by OPAL and
is shown in Fig. 7 [13]. Here the quantity R, defined as the ratio of the
measured hadronic cross-section to the theoretical muon-pair cross-section
is shown as a function of the center of mass-energy. The two sets of data
points correspond to non-radiative events (open dots) and inclusive (all)
hadronic events(black dots). The non-radiative cross-section (and the the-
oretical f.L+ f.L- cross-section) were corrected to Born level, where Born level
means the improved Born approximation (see section 9.1) of the program
ZFITTER, which was used to compute the corrections. In the inclusive data
one can see the onset of WW and ZZ production, the latter very small and
only at the point of 183 GeV energy. The dotted line at energies above 160
GeV is the prediction without WW (and ZZ) production, which is clearly
required to explain the data. The Figure also contains low energy data from
PEP, PETRA and TRISTAN. At lower energies R is mainly R"{ and at the
Z peak it is mainly Rz. At the higher energies the "{ and Z contributions
are of similar importance.
(J (e +e -' >ff) LEP (preliminary) A (e+e'->ff) LEP (preliminary)

21) r_"""'T""''''~'''''~-'P~''''''''''''' 'T'''''''''T''''T~'F'''':: 41)
15 f
20
III
10
-10
-IS
.
• e+e ....> ~l+-

~
II
-2.
• e+e --->11+--
~l
1j
:: L. . .,. . . .__
-20
HID 182 lK4 186 lK8 190 192 194 1% 19K lHO
sqrt(.' i,(:t!V
180 182
L~
L.,.. d .........L •..••..
184 18fi lKH 190
•.•..L •." " .. L" ....... [ .•..•.•. 1.. ....... ..
In 194 196 19M 200
sqrlCs)
Figure 6. The e+e- annihilation cross-section and the leptonic forward-backward charge
asymmetries for non-radiative events (8' /8 > 0.85) minus the Standard Model expecta-
tion, for the four LEP experiments combined [15]-
These events have also been used to study several topics, namely:
1. measurement of the "{ - Z interference,
2. measurement of the energy-dependence of CY.qed,
3. limits on contact interactions,
4. limits on interactions mediated by new heavy particles,
8 E. FERNANDEZ
OPAL • Inclusive
,,"j Bom
20
IS
10 . .~.*" .
.......~ .. "*..." .....
+ PEPIPETRA
TRISTAN
• TOPAZ 95
020 40 60 80 100 120 140 160 180 200
sqrt(s) /GeV
Figure 7. The ratio of the measured hadronic cross-section to the theoretical 11+11-
cross-section. The open points correspond to the cross-section for non-radiative events,
corrected to Born level, while the black points corresponds to inclusive events. The points
at the peak and the theoretical 11+11- cross-section are also corrected to Born level (see
text and ref. [13]) .
which are covered below.
3.1. MEASUREMENT OF THE I'-Z INTERFERENCE
For non-radiative events the 'Y and Z diagrams have similar amplitudes
and offer the opportunity of measuring the 'Y - Z interference term (which
was fixed to the 8M value in the electroweak analysis at LEPl, where it
is very small). Fig. 8 shows the OPAL measurement of the quantity jho;d'
which measures the hadronic Z - 'Y interference (see ref. [12]) , versus the
Z mass for events at the peak, and including LEP-2 data up to 172 GeV.
The inclusion of the high-energy data considerably reduces the error with
respect to LEP 1 alone.
3.2 . MEASUREMENT OF THE RUNNING OF (XEM(S)
For non-radiative leptonic events the dominant diagram is 'Y exchange in

the s channeL The corresponding cross-section is proportional to (X~m'
For hadronic events the dependence on (Xem enters through both the
photon exchange and the vector part of the Z exchange diagrams, but the
effects tend to canceL
OPAL has made a fit to the data, namely to the cross-sections for lep-
tonic and hadronic final states, forward-backward asymmetries for muons
...1 OPAL <> .......... LEP 1
• -LEPl+lJO..l72GeV
0.75
0.5
-0.5
-0.75
·1
91.17 91.18 91.19 91.2 91.21 91.22

mz/GeV
Figure 8. Measurement of the hadronic "f - Z interference from OPAL (see text).
and taus, and Rb (the ratio of the Z width into bb to that into qij) [13] .
.__150 rr-r-rrr-r-rr~rr-r"""""-"""""T""T"T-,-r-r..,.,-r""T""T""T"""""""'-'-'
o o OPAL fit to (Jrr AFB and Rb
'"7'-' 145 ..... TOPAZ and OPAL averages
~
*: <> 0 l'> Fit to (In and AFB
140 a-I(O)
135
130
TOPAZ
125 AMV
TOPAZ
120 VENUS
115
110
105 assuming SM running up to a-1(QIUml
100 0 20 40 60 80 100 120 140 160 180200

Q/GeV
Figure 9. OPAL measurements of (tern together with measurements at lower energies

[13].
10 E. FERNANDEZ
Two analysis have been done:

(1) Fixing the other SM parameters (in ZFITTER). The result is
a;~(157.42 GeV) = 119.2:!=g
The quoted y'S value is the luminosity-weighted average of the data sample.
This result depends (through the luminosity) on the assumed running of a
from Q2=0 to typically Q2 = 3.5 GeV 2.
(2) Using measured values of a(p,+p,-)/a(qq) and a(T+T-)/a(qq)
The result is
a;~(157.42 GeV) = 119.9::~ :~ ± 0.1
which is independent of the running of at low Q2. This value differs by 2.6
standard deviations from the value at Q2 = o. OPAL has also combined
their measurements with those at TRISTAN. The result, extrapolated to
the Z, is
a;~(Mz) = 121.4:!=tg ± 0.1
(see Fig. 9). This result is statistically-limited and does not depend on as-
sumptions about the running of a at low Q2, which is the main uncertainty
on the value
a;~(Mz) = 128.90 ± 0.1
used at LEPI for electroweak analysis.
3.3. LIMITS ON FOUR-FERMION CONTACT INTERACTIONS
New interactions at a much higher energy-scale A, can show up at lower

energies in extra terms in the SM Lagrangian
g2
L = L
(1 + 6)A2 i,j=L,R
7Jij[en lle i][!nll!j]
where eL,R and h,R are the left- and right-handed spinor projections, 7Jij
depend on the model and A is the energy scale of the new interaction. A
can arise from the exchange of a very heavy particle, or can be due to
substructure of the fermions, or, generally speaking, can be considered as
a parameterization of new physics.
The consequence of these terms is a dependence of the differential cross-
section on:
E = (1 + 6)A2
d~
--{) = ~SM(S, t) + C20 (S, t)€ + C40 (S, t)€ 2
dcosu
where the C depends on the specific form of the extra interaction term.
Figure 10. 95 % confidence limits of the energy scale from contact interactions. The
bars for each channel correspond to different models, indicated at the bottom.
Several models have been assumed and fitted to the data (after correc-
tions, including e-w radiative corrections). The lower limits on their value,
from OPAL [13J, are shown in Fig. 10.
3.4. LIMITS ON EXTRA Z BOSONS
Extra Z bosons are present in many GUT theories, such as E6 (several

varieties) and L-R models (SUc(3) Q9SUL(2)Q9SUR(2) Q9Uy(l). A se-
quential Z, (same couplings as Z but higher mass) can also be present in a
simple extension of the SM. For the latter, an ALEPH analysis [5J gives:
12 E. FERNANDEZ
M~(sequential) > 898 GeV (at 95%)

an less restrictive limits within several model scenarios.
4. Measurement of the WW production cross-section

One of the main goals of LEP-2 is the study of WW production, which
involves the triple gauge boson couplings Z - W - Wand "( - W - W
(see refs. [16] to [25]). This requires measuring the cross-section for WW
production. That is, we want to select events
where the W's are "real" (resonant). There are 3 main diagrams (called
CC03, Fig. 11) which contribute to this channel:
e'
Figure 11. The CC03 diagrams for WW production in e + e - interactions.
Each W then decays to two fermions, leading to a 4-fermion final state.

However, for any specific 4-fermion final state, there are many other
ways in which it can be produced, which do not involve the above 3 di-
agrams. For example, for the specific channel e+ e- -+ udp,lIJ.L final state,
there are 10 diagrams (see Fig. 12):
We have to agree on what we mean by "signal events". The LEP col-
laborations follow slightly different procedures, but in essence all methods
consist on selecting events that mainly "come" from those diagrams, and
compute the "effective cross-section" for them. The effect of the other dia-
grams is corrected for, via Me simulation.
E.g., in ALEPH:
N obs - N baek - N 41
eeD3
O'ee03 = L
fee03
'K~ e
"
W
G
e " 3
2 2
graph 4 graph 5
2 2
graph B graph 9
2
graph 10
Figure 12. Diagrams contributing to the final state e+e- -t udl.LlJI-"
N 4/ = L[ E4/
cd13 MC MC]
- Ecc03(} cc03
where
Nabs = number of observed events,
Nback = number of events from non-like WW diagrams
N~§03 = number of events attributed to 4f, non-CC03 diagrams,

E'S = selection efficiencies (computed from MC).
14 E. FERNANDEZ
The "4f-CC03 correction", N~fo3 / L, amounts to about 1% at most. The

W's decay leptonically (Zv) or hadronically (qll/2). The events can therefore
be classified as:
(A) Fully leptonic:e+e- -+ Z+Z-vD,
(B) Semileptonic: e+e- -+ ZvqllJ2,
(C) Hadronic: e+e- -+ qqqq.
The selection methods are different for each case, which are also affected
by different backgrounds. In general, the procedure to select these events
consists on the following steps:
- Make loose cuts to get an "enriched" sample of WW events, and elim-
inate as much background as possible.
- Make a more refined selection to get Nobs for each of the final state
classes considered (inclusive or exclusive).
- Apply 41 - CC03 correction to get cross sections.
The main background comes from
Z(-y)Z where Z -+ Z+Z- or Z -+ qq.
- Z(gZuon) where Z -+ Z+Z- or Z -+ qq.
- ZZ (above ZZ threshold), where Z -+ Z+Z- or Z -+ qq.
Radiative Z events have either a hard photon or missing momentum along
the beam direction, hence the cuts on these characteristics.
All the LEP collaborations have developed sophisticated methods of
selection for each of the different channels involved.
4.1. FULLY LEPTONIC W DECAYS
These events are characterized by

- two energetic and acoplanar leptons of opposite charge,
- missing momentum due to the undetected neutrinos,
- if one lepton is a tau, it can give rise to a narrow jet.
One then has 3 types of events:
- lepton-lepton,
- lepton-jet,
- jet-jet.
The events are selected by making appropriate cuts based on the above
characteristics. Examples of these variables are shown in Fig. 13 which is
the acoplanarity distribution of the two leptons of ALEPH, and in Fig. 14,
which shows the energy distribution of the leptons of L3. The arrows show
the position of the cuts.
1
:! ALEPH
'"
t,o' o
1
Me tWW-tlvlV + BIN:kgmllmJs,
MC(WW-+N"',
'0 2
Me {Blld:gmwuls}
/JArA
'0
.,
'0
30 90 120 150 l80

AcopIan.r,jly (~OOS)
Figure 13. Acoplanarity angle of the two charged leptons in purely leptonic decays of
the WiS .
p reliminary
>Q) (b) • Data Iv lv L3

C}75
n M.C. signal
138 M.C. background
0
<Il
C50
Q)
>
w
'0
~cut
0525
.0
E
•
:::l
Z
20 40 60 80 100
Elepton [GeVJ
Figure 14. Lepton energy in purely ieptonic decays of the Wi s.
4.2. SEMI-LEPTONICS DECAYS
These events typically have

one lepton of energy > 40 Ge V,
- large missing momentum,
- two hadronic jets, each of energy> 40 GeV.
An example of this kind of event is shown in Fig. 15.
16 E. FERNANDEZ
Figure 15. An ALEPH WW event where the W- decays leptonically and the W+
decays hadronically.
The selection criteria for lvqq are different for l = e, p, or l = T. For

example in ALEPH the selection of lvqq(l = e, p,) events, starts with a pre-
selection which consists on finding the direction of the missing transverse
momentum and taking as the lepton candidate the particle with the highest
momentum anti parallel to the missing momentum. It is then required that
the particle has momentum greater than 15GeV Ic and that it is identified
as an electron or a muon. After the preselection, three quantities, lepton
energy, missing momentum and isolation angle, are examined, and a prob-
ability is then calculated, based on probability density functions for those
quantities, obtained from a Monte Carlo sample of these type of events. A
cut is then made on the event probability computed in this way, and the
event is selected or rejected. The distributions of lepton energy, missing
transverse momentum and isolation are shown in Fig. 16 for ALEPH. The
probability distribution functions are shown in Figs. 17.
For the TVqq events the criteria are different. A cut is first made on
global properties of the event and an explicit attempt is made to identify
the tau.
At the end one obtains the number of events selected in each channel
and the estimation (from MC) of efficiencies, cross-channel missasigments
20
c 11 >,t. -' ALEPH
~ 18 .Js= -16 1 CF;:V
:; 1 6 , el/q(~/I-wq(~ s elo:- (_t ion
• DATA
[ _-=-) qq F(THIA • DATA
c
<,
> [=:: TT ~ ORALZ LJ qq PriHIA
W 6 CJ :1 P rTH IA [:J Tr I'"O RALZ
=
rnm
c:J
E:3
ZZ P(fH IA
Zee PITH IA
C 'Cl rrnru yy PHOT02
C,J Ivqq I" ORALW 4 f
21 ',~,~
o L~l.T,~~~~
20 30 40 SO 60 70 80 10 20 30 40 SO 60 70
,
80
Lep t o n( +y ) Energy «>V) Missi ng P, (GeV)
• DATA
12 0 qq PYTHIA 1 1 pb- ' AL EPH I
c::J TT " ORAL:
../s=161 (;tV
c CJ ZZ PYTHIA
~ 10 arm ;;~P~gT~~ ! €L-'qq/ fa'q(~ s~ lec lion i
c:3
,+
lvqq hORALi 4 <1
~
1 ~I ! ~ 1
41 ii ~ ~c*,*1
21 +~ (!
, ( I~
o ~'~-.-.-f',..,.e~
1 :~l
·14 ·12 ·10 ·8 ·6 -4 ·2 2 4
Le pt on Isola ti o n
Figure 16, The distributions of lepton energy, transverse momentum and isolation
variable for semileptonic WW events, as measured in the ALEPH detector .
and backgrounds [18] (see the table of Fig. 18).
4.3. FULLY HADRONIC DECAYS
About half of the WW events decay in this way. They are characterized
by 4 separated and energetic jets. The main background comes again from
qq(-y), qq(g) and ZZ (above ZZ threshold) events decaying into 4 quarks
(a practically irreducible background). An example of such an event is in
Fig. 19.
Many methods have been developed to select this channel, combining
all the available information (event shape variables, invariant masses, etc.)
in an optimal way.
18 E. FERNANDEZ
• OATA ALEPH ( 0)
V'\ , ~ ~~~~. evqQ selo:·: hol1
C! 10 .... L:j 7.2. PYTHIA
o .... ~ Z« PYTHIA
..... IIC} We \I PYTHIA
C :::.:J .:vqqKORALW4(
&10 1
1 1
-,
10
o 0_1 0.2 0 .3 0 .4 0.5 0.6 0 _7 O _R 0_9 1 o 0.1 0.2 0.3 0.4 0.5 0_6 0_7 0_8 0 _9 1
Prob(evqq) Prob(j.lvqq)
Figure 17. The probability distributions for the selection of evqq and /-wqq WW decays
in ALEPH.
Evertse1ectmmc1assication
ee e)l e )l)l )l e qq )l qq qq ~q 11
ee 68.2 - 8.4 - -- - - - - 76.
e)l 0.1 70.8 2.0 - 2 . 9 0.3 - - - - 76.
e 4.3 3.8 57.9 - 0.1 3.5 0.3 - - - 69 . ~
E . fo )l )l - - - 71 . 1 4.7 0 . 2 - - - - 76 .
ww ! )l - 4 . 1 0 . 2 3.9 61.7 1.5 - 0.6 - - 72.0
(it%) - 0 . 7 5 . 1 0 . 3 5 . 6 45.2 - - - - 56 . 9
e qq - - - -- - 81. 2 0.3 6 . 2 - 87 .
)l qq - - - - - - 0.2 88 . 6 3 . 5 - 92.3
qq - - - - - - 3.1 3.3 54. - 60.
q~q - - - - - -0.1 - 0.1 84 . 84 .
Bad<grounds
0 . 020_010 . 050 . 01 0 . 02 0.0 0 . 11 0 . 05 0.1 1.2 1.6
(iIPb)
6 14 18 8 11 4 1127 113 861 43211 8181
Figure 18. ALEPH efficiencies and backgrounds for WW events at 183 GeV c.m. energy
(CERN-EP /99-035).
E.g., in ALEPH, at threshold [16], four methods were used to select

WW hadronic events. All of them start with a pre-selection (large visible
energy, small missing momentum, 4 jets and others) of efficiency '" 90%,
followed by a method of estimating the number of events in each channel:
Linear discriminant analysis :
Figure 19. An example of a WW event where both Ws decay into two jets as seen in
the ALEPH detector.
where Xi are variable which are different for signal and background
(Durham Y34, minimum jet energy and others). Events are then se-
lected by a cut in U.
- Rarity analysis :
For each event a number of variables Xj are calculated, such that the
value of Xj is small for background and large for signal. A new variable
T i , the fraction of Me WW events for which Xj < X} for all variables
j, is computed. The Rarity ~ is the integral probability of Ti, that is
the fraction of Me events for which T < T i . This distribution is then
fitted by a maximum likelihood method.
- Neural Network:a neural network incorporating 19 [18] or 14 [44] rele-
vant variables (global event properties, jet properties, WW kinematics
and others).
- Event weights:
The cross sections are directly computed from the events, which en-
ter with a weight, calculated by Me in a multi- dimensional space of
discriminating variables.
These methods have been adapted for selection at higher energies. The
results for the 4-jet channel are also shown in the Table of Fig. 18 [18].
Once the number of events in each channel are known, together with
the expected background and efficiency matrix, the cross section can be
calculated, e.g., by means of another max-likelihood fit: the cross sections
20 E. FERNANDEZ
for all the channels j enter in the likelihood
where P is the Poisson probability of observing Ni events in channel i when

ni are expected. The ni are given by,
ni =L x (I: Eij(Jj + (J~ckg)

j
where L is the luminosity, (Jj the parameters we want to determine, and the
(J~ckg and efficiencies E are computed from the table of Fig. 18. One can also
derive the (Jj from the total cross-section and the luminosity, assuming SM
branching ratios for the W. In this case the total production cross-section
is the only parameter.
The latest results, combining the four LEP experiments together, were
also presented at last week Tampere conference [15] and are shown in
Fig. 20. The data clearly require the existence of the ZWW vertex, as
predicted by the SM.
189 GeV : preliminary
L
20 /
.0 /
a. • I
/
......... : I
.o ,
• I
~
.......-
I • I
:/
~
./
./
10
I
+(!)
I
(!) • Data
- - Standard MOdel
.......-
b - - - no ZWW vertex
_. . . . .. ve exchange
0
160 170 180 190 200
sqrt(s) GeV
Figure 20. The WW cross-section for the four LEP experiments combined, as a function
of the energy.
ALEPH
I
OPAL
LEP W5 ev
I
''i
.3'1 " 0 49
,.,
10.52 ± 0.45
10.61" 0.25
±
:t .; 4-'~
ALEPH
II
-- 66.89 ± 0.74
ALEPH 1129 ± 046 L3

L3
OPAL
- ..-~ 10.47 ± 0.42
±
J ~b 2: , j 4'3
•
OPAL 68 .34 " 0.68
....
LEP W5 PV 10.65 ± 0.24
ALEPH 10.48" 0.00 LEP 67 .96 ± 0.41
:
-
±
~~ _ I '0 ± Di,l!
OPAL -it- 10.69 ± 056
LEP W5 "TV 10.82 ± 0 .32
LEP W5 Iv 10.68 ± 0.13
10 11 12 66 68 70
Br(W- :>Iv) % Br(W- >hadrons) %
Figure 21. The W ieptonic and hadronic branching ratios for the four LEP experiments
together [15].
5. Measurement of the W width

From the above measurements it is also possible to extract the leptonic
and hadronic branching fractions of the W (e.g. from table of Fig. 18). The
latest measurements are shown in Figs. 21 , taken again from [15].
In these figures the prediction of the 8M is also shown. The calculation
within the 8M of the partial widths into a particular fermion-pair final state
is simple and given by
r W-tff' a Mw ITT 12 ( )
-_ Nf
Born
c -6-2- Vij F mf,mf,Mw
" J 2sw 'J
where sw stands for the sine of the electroweak mixing angle, Nc is the
color factor (1 for f=lepton and 3 for f=quark) and Vij is the element of
the Cabbibo-Kobayashi-Maskawa matrix for the flavor indexes i and j. The
function F is equal to 1 if the fermion masses are neglected with respect to
the W mass. For leptons Vij = oij. For quarks the possible pairs (fi, fj) are
(u , d), (u, s), (c, d), (c, s), and the strongly Cabbibo-supressed (u, b), (c, b) .
After summing over all of them the total width is
rBam _ 3a Mw
tot - 2 2s~ ·
Radiative corrections can be conveniently included by expressing the
width in terms of Mw and GJ1,' giving Improved Approximation (see section
9.1) widths:
22 E. FERNANDEZ
r lEA
w . . . . vi1j
r lEA
W ....... Uidj
Assuming Lepton Universality
Be = BJ.t = Br = (1 - B qq )/3
the data can be fit to the cross-sections (at each energy) and B qq .
The result can be put in terms of the CKM matrix elements
B(W -+ qq) frs(Ma,) 2 2 2 2 2 1 12

1_ B(W -+ qq) = (1+ 7r )(lVudl +lVcdl +lVus +IVcs lVubl + Vcb )
1 1
From the world average values of as frs(Mi) (evolved from frs(Mi) ),

together with other measured CKM matrix elements, a value for the least
well measured IV csl element can be obtained. The result at Tampere for
the four experiments combined was [15]:
IVcs 1 = 0.997 ± 0.020

An alternative way of obtaining this number is to look at charm decays
of the W. Since W -+ tb is forbidden by energy conservation and W -+ cb is
strongly Cabbibo-suppressed, it is possible to tag charm decays (predom-
inantly from W -+ cs) and measure the corresponding BR. Assuming the
unitarity of the CKM matrix this branching ratio is
This has been done, using various methods to select charm events, by all
the LEP collaborations. The latest results presented at Tampere are [15]:
The question of whether or not the W could decay into undetected par-
ticles (e.g., low momentum charged particle below detectability) has been
investigated by ALEPH [18] . The idea is that, in this case, the total width
would be modified with respect to SM expectations, and this modifica-
tion would affect the total cross-section in a small way, namely the visible
cross-section would become
F(W -+ cx)/r(w -+ had) Wesl

ALEPH 0.51 ± 0.05 ± 0.03 1.00 ± 0.11 ± 0.07
OPAL 0.47 ± 0.04 ± 0.06 0.91 ± 0.07 ± 0.11
where
rvis
=
B vis
rw +wrwvis = 1 _ Binvis
ALEPH made a fit to the measured cross-sections at 161, 172 and 183
GeV, taking the visible width from the SM and Mw from the world average
(excluding LEP).
The results were
rwvis 30~~~ (stat.) ± 33( syst ) Me V

rwvis < 139 MeV at 95% CL
Binvis < 6.5% at 95 C L.
6. Measurement of the tri-linear couplings VWW (V =" Z)

The triple gauge couplings in the SM (Fig. 22) are a consequence of the
non-abelian structure of the electroweak interaction.
W+(q)
V(p) ~ = --- iegvwwr v:6/>(q,q,p)
"\. g-yww = 1 .
W-(q) . 9zww = cot Ow
6z = gzww - cot Ow
Figure 22. The VWW vertex, where V is a photon or a Z boson.
The measurement of the WW cross-section clearly establishes their ex-

istence, as we have seen above. The exact form of the couplings also has an
24 E. FERNANDEZ
effect on the production angles of the W s and on their polarization (and

therefore on the angular properties of their decay products).
One would like to investigate if other couplings, different from those of
the 8M, are also present. The most general way of writing the cross-section
(compatible with Lorentz-invariance and U(l) gauge invariance) is in terms
of 14 couplings. Of the 14 only 5 preserve C and P and are likely to playa
role at LEP-2 (see refs. [26] to [36]):
(In the 8M gf = kz = k"( = 1, and all the others are 0).
e----~------~~~----+----
f
Figure 23. The angular variables used in the analysis of the triple gauge couplings.
The W - W - 'Y couplings are related to the magnetic-dipole moment

and the electric-quadrupole moment of the W:
LEP1 measurements constrain the deviation of the couplings from their 8M

values (they enter in LEP 1 observables through loop-corrections). To look
for deviations one tries to extract from the data
~gi = 1 - gf , ~k"( = 1- k"(, A"(

which should be zero in the 8M.
The relevant observables to measure the anomalous couplings are:
- Cross-section and production angles of Ws.
Polar angle of decay charged-lepton or d-type quark with respect to

the flight direction of the corresponding W in the rest frame of the W
(see Fig. 23).
Azimuthal angle (as above) with respect to the plane formed by the
W direction and the beam.
The reconstruction ofthese angles is different for semi-Ieptonic or hadronic
decays. In qqqq decays there is the ambiguity of which of the four jets be-
longs to the d-type quark. Overall kinematic fitting of the event is always
done, to improve the resolution on the angles.
300
t dltta: qqqq L3 t dlt,a: qqlv L3
300 nsignal
pdlmIll.'
W
n,ignal
JA bo<:kgrvund
~'" ,.)
JA background
~ <Iog!=+1 "~ <Iogt+1
,i ., <10&,--1
"~ ." d~81-= · 1
~2OO
C c
& ~
w
~
0 'Z
~5 .8E 100
1
Z ~
0., -0.5 0 O.S <>'1 .{l.S O,S

co....e w · ~ew·
200 t dlt'a; qqlv L3 t data: qqlv L3

,.)
..
ptdCmll'llry
n,ignal ~Ul lll"wy 200 Il.signal
nb,,,,tground (1)'
In background
" <Iog'=+1 . ' . <Iogj=+1
ISO ,:':, <Iog:=-I I;l ,: . <Iog,=. 1
1
150
>
OJ
'0 100
.ll8
SO i. 50
.{l,S 0
<>'1 O. O.S 2 4
CQ~e 
Figure 24. The distributions of the angular variables for the analysis of the trilinear
couplings obtained by the L3 collaboration.
The LEP collaborations have tried several methods of fitting:

Maximum likelihood: the probability of occurrence of an event with the
measured variables at each energy can be calculated by Me simulation,
taking the measured cross-section for the corresponding energy. The
problem with this method is how to introduce background and detector
26 E. FERNANDEZ
effects. This problem can be avoided with a binned likelihood method,

but the statistics is very limited to proceed in this way.
- Another technique is that of the "Optimal Observable". For a given
anomalous coupling O:i, the cross-section can be written as
da· . . 2
do' = CMn) + Cl (n)O:i + C2(n)O:i
The quantity CI/Co (computed after folding-in ambiguities) is the optimal
observable (00) for O:i and can be computed for each event j. A maximum
likelihood fit can then be performed to obtain O:i:
n
InL = L InP(O~ ,O:i)
j=1
To show the sensitivity of the angular variables to anomalous couplings

the results of L3 at 189 GeV are shown in Fig. 24, and the results from the
four LEP experiments, together with the form of the likelihood functions,
in Fig. 25 [37].
ALEPH + DELPHI + L3 + OPAL

3 3
~
0 ::!.S ::!.S
<? ::! ::!
1.5 1.5
I 1
0.5 0.5
0 0
-{).4 -{).::! 0 O.::! 0.4 -D.::! -{).I 0 0.1 O.::!
~K.., t.~
3 +0 .09
~
0 ::!.S A'Ky= 0.06 -0.09
<? ::!
+0.03
1.5 Agr= -0.01 -0.03
0.5 +0.04
D
Ay = -0.03 -0.04
-D.::! 0 D.::!
\
Figure 25. The values of the anomalous couplings resulting from the four LEP exper-
iments. To obtain each value it is assumed that the other two couplings have the 8M
value. Each of the curves in the plots is for a different LEP experiment.
7. ZZ production
The 183 GeV and 189 GeV center of mass energies are above ZZ threshold.
The data have very recently been analyzed and new results were presented
at the Tampere conference last week [38]. The analysis parallels that of
the WW. Again what is observed are 4 fermion final states coming from
the decays of resonant (quasi-real) Z's. The relevant diagrams (NC02) are
shown in Fig. 26.
Figure 26. The NC02 diagrams for ZZ production in e+ e- collisions.
~
'.
,
Z.Y
z.Y
,
f '.'
r
<
~
Z.Y r
~z.Y '"
...r
~ c·
Figure 27. The Non resonant diagrams contributing to a 4 fermion final state.
But for every 4-fermion final state the events can also be produced by
non-resonant Z's or /, as in Fig. 27. Thus, as in the WW case, the events
are selected with the expected characteristics from two real Z decays, and
the effect of the non-resonant diagrams is estimated from the Standard
Model and included as a correction. Fig. 28 shows the results for the four
LEP experiments at 189 GeV, while Fig. 29 shows the average of the four
experiments at 183 and 189 GeV. The curve in Fig. 29 is the Standard
Model expectation.
28 E. FERNANDEZ
Tampere 99 - Preliminary - 189 GeV
(jZZ (pb)
ALEP H O 67 ';' 0.14

· ·0.14
l3 O 75 .;. 1)· 15
· - r' 11
OPAL 076 + 0.16

· ·0.1 5
LEP 0.70 ± 0.08
common 0.015 pb
·lldof = 0.63 / 3
SM 0.65 ± 0.01
I ••• • I .---t- • • I •• , , I
o. 1.
all (pb)
Figure 28. The ZZ production cross-section at 189 GeV [38].
8. The measurement of the W mass and its implications

One of the major contributions (if not the major) of LEP-l to particle
physics has been the verification of the electroweak theory at the quantum
level. This has been possible once the properties of the Z boson, it mass,
width and couplings to fermions, were measured with precision. In the
Minimal Standard Model (MSM), which contains the Electroweak Theory,
any observable, 0 , can be calculated as a function of a few parameters,
which, for analysis of the Z , can be conveniently taken as
where a(Ml) is the electromagnetic coupling constant at the Mz scale, Gil

is the Fermi constant, Mz is the Z mass, mj are the fermion masses, MH
LEP - preliminary
I.S , - - - - - - - - . ! . - - - - - - ' ' - - - - - - - - - - ,
t' Data
_ Standard Model
0 .5
175 180 IR5 I~O 1~5 200

Ecm(GeV)
Figure 29. The ZZ cross-section as a function of the center of mass energy [38) .
is the Higgs mass and, O:s is the strong coupling constant. That is,
The dependence on the fermion and Higgs masses and on O:s enter only
through radiative corrections. The precise measurements of "electroweak
observables" at LEP-l , together with other precise measurements obtained
at hadron colliders and in high energy neutrino scattering experiments, can
thus be used to test the consistency of the MSM, or, if the MSM is assumed,
to constrain its parameters (see below).
With the set of parameters mentioned above, the W mass is a derived
quantity. It is thus important to compare its experimentally measured value
with the calculation based on known values of the other parameters, as
measured in LEP-l or elsewhere. At the present level of precision obtained
at LEP-2 and at hadron colliders [39], it is necessary to include in the
calculation non-trivial one loop corrections of the SM. Alternatively we can
consider the W mass as a fundamental parameter of the Standard Model ,
and use it as such to constrain, together with the measurements of LEP-l,
the value of the yet unknown Higgs mass.
30 E. FERNANDEZ
In the next subsection a very brief remainder of the LEP-1 measure-

ments is presented. This is followed by the explanation of the measurements
of the mass at LEP-2 and its comparison with the SM predictions.
8.1. PRECISION ELECTROWEAK MEASUREMENTS AT LEP-l
One of the (pleasant) surprises of the early analysis ofLEP-1 and SLC data,
was the realization that it was actually possible to do measurements with
higher precision than anticipated in all the previous studies [1]. The very
precise knowledge of the LEP energy, together with the precise measure-
ments of the luminosity at the four experiments, and the very clean classi-
fication of the final states in specific lepton-antilepton or quark-antiquark
pairs, have permitted, for the first time, to test the Standard Model at the
one-loop level radiative corrections. There are many reviews of this subject,
both from theoretical and experimental points of view (see for example [40]
and references therein).
There are three major types of "electroweak observables at LEP-1:
(A) Lineshape{s) (and from them the mass and the total and partial widths
of the Z).
(B) Forward-Backward Charge Asymmetries.
(C) Polarization Asymmetry (for 7+7- final states only).
These quantities, and others related to them, are (almost) directly measured
from the data. In a further step, these observables are interpreted within
the Standard Model to test its consistency and/or constrain its parameters.
This is a program that started at LEP 10 years ago and which is about to
finish. The quality of these measurements is likely to remain unsurpassed
for quite a few years, since no accelerator is foreseen at the moment where
they could be improved.
The lineshape for f J is the cross-section as a function of the energy of
the reaction:
At tree-level it is easy to calculate the cross-section from the two Feynman

diagrams of Fig 1. The cross section has 3 terms, corresponding to the
exchange of the 'Y, the exchange of the Z, and the interference between
them.
{For the e+ e- -* e+ e- case there is the additional t-channel photon ex-

change diagram, which has to be added, or, alternatively, the data have to
be corrected to "effectively remove" its contribution. It is not considered

any further here).
The, annihilation diagram gives
_ 47f0:2 N 2 Q2
{J'Y - 3s c I
where 0: is the fine structure constant, Nc is the color factor (1 for leptons,
3 for quarks), and QI is the electric charge of the final state fermion. The
interference term is very small at the peak and can be included as a correc-
tion, calculated in the SM. The term due to the resonant Z exchange can
be parameterized, in a model independent way, as a function of the total
width of the Z, fz, and partial widths fe, f I, of the Z into electron (from
the production) and f pairs(from the decay), respectively,
11 _ 127ffefl sf~
{Jz (s) - M2
z
r-
(s _ M2)2 + M2f2
z z Z z
where the widths are given in terms of the vector and axial vector weak
neutral current couplings gv,f and ga,f, respectively. We will write them as
gvl and gal for simplicity.
At tree level the widths, and hence the line-shape into any particular
f 1 final state, can be expressed in terms of just 3 parameters: 0:, Mz and
the Fermi constant GJ.t which are very well determined experimentally (Mz
measured at LEP). They are given by the following relations:
_ NjGJ.tM! 2 2. "
rl- 6V27f (gvl+gal) , rz=~fj
f
In the S.M. and at the tree level
gvl = (If - 2QI sin2 Ow)

gaj If
sin2 Ow = 1- Mar
M~
GJ.t =
That is, at tree level the widths are functions of 3 parameters
Notice in particular that Mw is also given in terms of these 3 parameters.

32 E. FERNANDEZ
With the level of precision at which the lineshapes (and hence the
widths) and other observables (see below) are measured at LEP, the tree-
level relations are not adequate to describe the data, but it is necessary to
introduce higher order radiative corrections. These corrections have been
studied in detail for the past 15 years. They can be divided into photonic
(or pure QED) and "non-photonic" (or electroweak) corrections. The most
important photonic correction is that due to initial state radiation. It can
be included by convoluting a radiator function (probability of radiating a
certain energy from the initial state) with the cross-section at the corre-
sponding reduced center of mass energy. They have the effect of displacing
the peak cross section by about +90M eV from the Z mass and of lowering
the cross section at the peak by about 30%.
Of the electroweak corrections the most important are those correspond-
ing to vacuum polarization diagrams (they are also called "oblique correc-
tions") They have a property called non-decoupling: masses much larger
than those of the Z show up in the corrections and do not vanish. In par-
ticular the top and the Higgs masses give contributions proportional to the
square of the mass, mr, and to the logarithm of the Higss mass, log mH.
There are several schemes to introduce the electroweak corrections. A
particularly convenient scheme for LEP-l is that of effective couplings: the
radiatively corrected cross-section is written in terms of the widths as in
tree level (with a modification consisting on replacing the width term in
the denominator, M~r~, by a "s-dependent width", (8rz/Mz)2), and the
widths are also given in terms of the coupling constants as in tree level, but
the coupling constants now become "effective coupling constants"
0(8), g;,j(8), g;,,(8), sin2 Oeff(8).

These are complex quantities which depend on the energy scale 8 (8 = M~
in the case of running at the Z peak, see section4.2) and , via radiative
corrections, on the fermion masses, in particular mtop and M Higgs, in the
form mentioned above. Light fermion masses, the electromagnetic and (for
quarks final states) the strong coupling constants at the Z scale also enter.
The relevant formulae are
(7
o 8r~
ff (8 _ M~)2 + (~)2
12rr rer j
M~G
#(1 - 4sin Otff(M~))
2
#
mz 91187.2 ± 2.1 MeV

fz 2499.4 ± 2.4MeV
a~ad 41.544 ± 0.037nb
Re 20.803 ± 0.049
R/L 20.786 ± 0.033
RT 20.764 ± 0.045
AO,e 0.0145 ± 0.0024
FB
AO,/L 0.0167 ± 0.0013
FB
Ao,T 0.0188 ± 0.0017
FB
TABLE 2. The basic LEP

electro-weak observables, as of
July 1999 (15).
The p parameter is not equal to 1, as at tree level, but is affected by

radiative corrections and depends in principle on the final state fermion f.
The electroweak mixing angle is not related to Mw and Mz as in tree level.
It becomes sin 2 () e j j, and it is also a function of the scale s and the final
state fermion f. (The reader is referred to [40], [41] for detailed definitions
of the effective couplings and their explicit dependence on the fermion and
Higgs masses, as well as the inclusion of more refined corrections, such as
final state photon and, in the case of quarks gluon, radiation).
In addition to the cross-section there are also other quantities that are
measured at LEP and are sensitive to electroweak corrections. These are
the forward-backward charge asymmetries and the forward-backward po-
larization asymmetry for T+T- final states.
The F -B charge asymmetry is the difference between the cross-sections
in the forward and backward hemispheres for a given f J final state, nor-
malized to the total cross-section for that state. An event is in the forward
hemisphere if the angle of the fermion with respect to the incoming e- (or
the angle of the antifermion with respect to the incoming e+) is less than
90 degrees,
Ai _ O"F - O"B
FB - O"F +
O"B
In terms of the effective couplings, and at the Z peak, the asymmetries

are
Aj ~ 29ve9ae 29vj9aj 3A A
FB - 4 -2 + -2
gve
-2+ -2
gae gvj gaj
= 4' e j
34 E. FERNANDEZ
Tot.1 width r 1.
., , ,.
111.11'111 - 0 - - 1)1'1 ." 111 ~IJ ,' .l:K O.II6{1
'-' -/!r- l!'OI.7 .!. 4 , IMeV 103 llUlt.l ::t: 0,061
01'\1 lll"t.
.20,'.0.11: 0.0::01
,GOO
"'0<1
... "".t co,". 2.1Mt\'
lhl,,' . 'U/l ...
~ ...
...
'~
ml, =9 J 187 1" 2 Mt:\'
E- El m L = .,.a ± S(;otV
O"=,-o,l l" III .. !} O'/"'I t II'Jf1.flU'l

"'" 10.'
Prelim inary
~bO.1 0.23107 ± 0.00053
CJ' .... A 0.:23:210 ; 0.00056
Au 0.23136': GOOO;;;!;;
~bO.b -.- 0.23228 ± 0.0 0036
PFI .I'HI ~bO.< - -. - - 0.23255 ± 0.00086
<O,u> 0.2321 . 0.0010
u 41.S.l{, ~ IJ_U.s~ 1111
Average(LEP) -0- 0.23192 ± 0.00023
l?Id.o.f 4.a , s
I .U' 41.54-1 · u.tI.}7nh Average(LEP 0. 23 151 ±0.OOOI7

l/d.o.t,: 11 ,9/6
10.0
10
... fttl l t ..m ' .'l1l11b

X1,,, ... . . . "J
J:
g r:n.. .. 1i4 :t 5C;I!'V
E 10
lOO
0.23 0 .232 0 .234

.1....t 41..5
O~ l nh l
"1.6
e
. 2 lept
sin elf
Figure 30. The Z width, the ratio of the hadronic to the leptonic widths (assuming lepton
universality), the hadronic peak cross-section and the effective electroweak mixing angle
from the LEP experiments. The sensitivity to the Higgs mass is shown [41).
where AI is defined as
29vi9al
9;1 + 9~1
The chirality of the weak current (it only affects left-handed particles
and right-handed antiparticles) and the approximate conservation of he-

licity at large energies, causes that the final state fermions are polarized,
despite the fact that the incoming e+ and e- are not. In practice the polar-
ization can only be measured for 7+7- events, since they decay inside the
detector and the angular distribution of their decay products depends on
their polarization, again due to the chirality of the weak charged-current
responsible for the decay. Two quantities are measured at LEP: the aver-
age longitudinal polarization, PT , and the forward-backward polarization
asymmetry. They are given by
(T~t (s) - (T~t

(T~t(s )
((Tr (s) - (T~ (s )) - ((T~ (s) - (T~ ( s ) )

(Ttot (s )
where + and - refer to the 7 helicity.
At the Z pole
PT(M~) -AT
3
A~POL(M~) --A
4 e'
These quantities are not entirely independent of the Forward-Backward

charge asymmetry: at the peak the tau longitudinal polarization as a func-
tion of the angle (between the final state fermion and the electron beam)
is given, at the Z peak, by
P ( 0) __ AT (1 + cos 2 0) + 2AecosO
T cos - 1 + cos 2 0 + 2AeATcosO
Thus the "Electro-Weak Physics Program of LEP 1" goes as follows:
1. Measure cross-sections for e+ e - -t f j as a function of the beam energy
(the "lineshapes"). For this, one needs to:
- measure the LEP energy,
- count number of events in the different final states f J,
- measure the luminosity.
From the last two steps one gets the cross-sections. These three quan-
tities have been measured at LEP (by the experiments and by the
machine team) extremely well, in fact better than foreseen. The errors
on the cross-sections are now below 0.5 percent, and the error in the
LEP energy is below 0.5 per mil.
36 E. FERNANDEZ
0.022 ,-,.......,.........--.--,--...-.,.-,--,-.-:P'r'el'
im'in- a-rry' -0.031 , -..--r--.--.--r-.....,---.-.....-,-,....P...,.r-el....
im'I'aTry'
........... 68% CL
"
I
................,r....... .
\" '~\
0.018 -0.035
'>
Cl
A,(SLD)
-0.039 ..............
0.014 mH
)
....................................
-rr
.... e' e
-11 P
.
. ·-- t +t -
68% CL
-0.043
20.7 20.8 20.9 -0.503 -0.502 -0.501 -0.5
gAl
200 .---------.-~~~~~
m,~ 174.3 ± 5.1 GeV .. - LEP Data, GF • d
0.233 m,.= 95 ... 1000 GeV ............:
-All Data ....
@ //
.....
0.2325 180
(./..........................
·.u.....·····
0.2315
140
0.231
Preliminary 68% CL Excluded Prel im inary
83.6 83.8 84 84.2 10

r, [MeV]
Figure 31 . 68% contour probabilities for correlated variables. The top two plots show
the measurements for individual leptons (dotted lines) and assuming universality (solid
curves). The SM predictions for Mz = 91.1867 GeV, mt = 173.8 GeV, mH = 300~~~g
GeV, and a.(M~) = 0.119 ± 0.002 are also shown (the intersection of the arrows). The
arrow point in the direction of increasing values of mt, mH, and a •. The bottom left plot
comes from the LEP-1+SLD measurements. The shaded area is the SM prediction for
the values of mt and mH indicated in the figure. The point with the arrow shows the SM
prediction if only photon vacuum polarization corrections were included and the length
of the arrow represents one standard deviation in a em . To explain the measurements the
electro-weak non-photonic corrections are clearly required. In the bottom right plot the
dotted is what it is expected from LEP data alone, while the solid curve includes the
CDF/DO measurement of mt [41].
2. Measure asymmetries, forward-backward for leptons (p, and T) and

quarks (when possible, e.g. for b quarks).
3. Correct data for initial state radiation and derive other quantities such
as Re,RJ.L and R T , where Rl is defined as the ratio of the Z partial width
into lepton 1 to the partial width into hadrons.
4. Fit the data to obtain other derived quantities, such as the Z mass,
total width and coupling constants.
5. Study the consistency of the data with the Standard Model interpre-
tation or, if the data are interpreted in the Standard Model, derive the
parameters of the latter.
It is not my purpose here to cover the above program which is described
in many places. There is a working group formed by the four LEP collabo-
rations, the "LEP electroweak working group" , to average the data between
the experiments and to fit them (see ref. [41] and references therein). The
group has been extended to include the SLD collaboration at the SLC. At
the SLC the electron beam can be polarized and this allows the measure-
ment of two other quantities: the left-right asymmetry and the left-right
forward-backward asymmetry,
At the Z pole:
ALR(Mi) = Ae
-j 2 _ 3
AFB(Mz ) - 4Aj
Each of the LEP experiments has collected more than 4 million Z events in
the period 1989-1995 and the SLD collaboration at the SLC has collected
more than half a million events. Shown in Table 2 are some of the numbers
from the Tampere conference, and in Figs. 30, 31 some of the plots from the
electroweak working group [41]. One of the intriguing results that comes
consistently from the data is that the inferred Higgs mass is low. This is a
good reason to hope that the Higgs will still be found at LEP.
8.2. MEASUREMENT OF THE W MASS
This is one of the most important measurements in LEP-2 and it has been
extensively studied by all the LEP experiments [42]-[52]. The measurements
38 E. FERNANDEZ
of the W-mass starts with the selection of WW events as explained section

in section 5.
Once we have these events, there are three different methods to measure
the mass:
1. Measurement of the end-point of the charged lepton spectrum in semilep-
tonic W decays. This is a classical method of measuring masses. It is
clean and easy to interpret, but limited statistically, and we do not
describe it further.
2. Measurement of the WW cross-section near threshold. Near the WW
threshold the WW production cross-section is sensitive to the value of
the W mass. The advantage of this method is that all decay modes can
be used. But one needs to run very close to threshold, and therefore
the statistics will be limited and the background large.
3. Direct reconstruction. The mass can be directly obtained from the
reconstruction of the W decay products. This is the most appropriate
method above threshold. It uses most of the decay channels, but for 4
jet events there are soft QeD effects that have to be taken into account.
For all the methods one needs the beam energy, which is used as a con-
straint.
8.2.1. W-mass from the WW threshold cross-section.

The sensitivity of the cross-section to the W mass is maximum near thresh-
old. The optimal point is at 161 GeV center of mass energy, which was
chosen as one of the points for running LEP-2 because of this reason.
It is not simple to write the WW production cross-section as a function
of the energy since there are quite a few effects that have to be taken
into account. What we want to compute is the Ce03 diagrams of Fig. 11,
followed by the decay of the Ws. The decay of the Ws is introduced by the
convolution [53],[54]
where the ps are Breit-Wigner functions for non-zero width Ws,
1 Mwrw
Pw(s±) = -I
7r s± -
M2w +'M r 12BR
t W w
and where the s are the invariant masses of the internal Ws, and BR the
branching ratio if considering a specific final state. The ao inside the integral
is the cross-section corresponding to the ee03 diagrams of Fig. 11 and their
interference. It is given by the expression
ffiW from (jww at 161 GeV

..fs = 161.33 ± 0.05 GeV
,........7
.D
0.. ()\' ',\ = 1.t,!);[ ~~.;~ '"'
'-'
~ 6 11." ""' ?;t).40 '" (i, \-
.2 LEP Average
tl
<0
</)
Vl
5
eu 4
Vl
~ 3
~
079 79.5 80 80.5 81 81.5 82

row (GeV)
man
Fi All] LEP 161 GcV W
LEP EW WOf'klnl Group
Figure 32. The W mass from the WW threshold cross-section.
a occ03 (sO ,+,-

s s) =
where the c's are functions of the coupling constants and propagators indi-
cated by their labels and the Gs depend on the kinematics.
The radiative corrections to these expressions are difficult to introduce,
mainly because they do not factorize into photonic and weak corrections
as in LEP-l. Adequate approximations exist nevertheless to introduce the
main effects. These are
40 E. FERNANDEZ
- Initial state radiation. They are included by introducing a "flux func-

tion" (or an structure function) which plays a role analogous to the
radiator function used in the analysis of LEP-l data.
- Coulomb singularity. At threshold the two Ws are produced almost at
rest and move slowly. The approximation that they are free particles
is no longer adequate since they are affected by the long range elec-
tromagnetic interaction. This is included as a correction consisting on
the change
agc03 (s) -+ C8 c03 (s)(1 + oc(s, s+, s-))
where oc is a function of the kinematics. At 161 GeV it represents a
correction of the cross-section of 5.7 %.
- Improved Born Approximation. The treatment here is again more com-
plicated than at the Z, since the W mass appears both in the matrix
element and in the phase space factor. The so-called fixed width scheme
is used [41].
There are other final-state effects, namely color-reconnect ion and Bose-
Einsten correlations, which affect the cross-section but which are not im-
portant for the measurement at threshold. However, they have a substantial
effect in the direct determination of the mass from the jet reconstruction
and are described in the next section.
The dependence of the WW cross-section, at the fixed energy of 161
GeV in the center of mass, as a function of the W mass is shown in Fig.32.
The band in that figure is the measured cross-section averaged over the
four LEP experiments and the resulting one-sigma value of the W mass.
The values are also shown in the plot.
8.2.2. W-mass from direct reconstruction.

The mass can also be reconstructed directly from the kinematics of the W
decays. There are three different "channels" to calculate the mass:
(A) Leptonic: W+W- -+ Z+I/Z-I/.
In this case the system is underconstrained (due to the 2 neutrinos)
and the precision on the mass is small.
(B) Semi-Ieptonic: W+W- -+ ll/qq.
Here events are selected where one W decays semileptonicaly the other
in qq. These events are typically selected with high efficiency (80% for
e and f..l semileptonic decays, 40% for T) and low backgrounds (less
than 5%). The mass resolution is improved by imposing, through a fit ,
energy-momentum conservation (a 2-c fit in this case).
(C) hadronic: W+W- -+ qqqq
In this case both W's decay to qq, and the final state is a typical 4-jet
event. As one has to avoid the "non-resonant" and "radiative return"
Data at 189 GeV

qqllv
qqev
DELPHI
.~ JO . - . . - -~~ ...... ~-- " ---~-

J~
> . • ALEPH Prcllmm:uy ~ . 11lII.6GcV
r . ...
-
90 "'" - " ...
~ 10 ~ ~l::::'-'.m.1" " ~ ao
~
l
~
60
!O
- -'«' . •
. • I'O _"'..... ~
",,,,"'11:'1 ~
,
.1
~
(j
10
60
t _
>~ -
... ~ ~~l ..
," •• _ _ ...
~
~ ~
JO '
'0 f
", '
l
", : ,. ;
qqqq
prelimina ry
L3 :; ~
OPAL Preliminary,
,---:.;..:.:..:.-----".:...-.------,
~189 (;ct:V
(e) -
t'M WWS ~qqq ~
M C (W W"",,,)
. .....t.H~\' u.
Figure 33. Effective masses for different final states from the four LEP experiments [15].
backgrounds the efficiency is smaller than for the semileptonic case

(70% typically) and the background higher (15% typically). However,
the statistics is high.
Here a 4-c fit can be imposed to improve jet energy and angular reso-
lution, and thus mass resolution. Other fits can also be imposed (e.g.
equal reconstructed masses on both jet pairs).
An additional problem in this case comes from jet pairing. The four
jets can be paired into two di-jets in 3 different ways. The way this
is handled (e.g. in ALEPH [43]) is by choosing as the right combina-
tion that which gives the smallest different between the two masses
reconstructed from each di-jet, unless this combination has the sum of
the two di-jets opening angles, in which case the combination with the
second smallest mass difference is selected. At the end a comparison is
made with a Monte Carlo sample, treated in the same way as the data,
and this decreases the importance of the exact procedure followed to
42 E. FERNANDEZ
select the events.

The invariant mass distributions reconstructed from the 4 fermion final
state should have a Breit-Wigner form, which is however distorted due to
many effects, such as initial state radiation, detector resolutions, wrong
assignment of particles to jets and others.
Mw (GeV) non-4q Mw (GeV) 4q
ALEPH 80 343 ~o 098 ALEPH 80501'0121
d0224 Q' ·J5 L~. :-11-

,
OPAL 80.362 0.105 OPAL 80.345 0.134
± 0.063
t
LEP 80.313 LEP 80.429 . 0.0 89
FS' 0.058 G.V
LEP 0 ,017 G.V LEP 0 017 GeV
Z2/dot . 17 9/20 X'/OOl :IE 17_9/20
••• , • • • • - • •• • , I
80 .0 81 .0 80.0 81 .0
Mw (GeV) Mw (GeV)
Figure 34. The W mass for four-jet final states and for other but entirely hadronic final
states [15].
All the LEP experiments have developed methods to handle this prob-
lem. The most common is to compare measured di-jet mass distributions
with MC generated samples with different masses, and choose as the value
of the W mass that used to generate the MC events that best resemble the
data. The problem is how to avoid the generation of many Monte Carlo
samples. The procedure is that of the "re-weighting technique" which goes
as follows: a large sample of Monte Carlo events is generated at a given
reference mass, M~f. This sample of Monte Carlo events is used again and
again, each time with a weight for every event given by
where M represents the matrix element of the CC03 diagrams at the given
value of Mw and fw and the p{ is the four momentum of fermion j (of jet
in the quark case) of the four-fermion final state for event i . The W-mass
distribution of the samples are compared with the data until the best match
is found . The corresponding Mw mass is the best estimate of the real W
mass.
Technically one makes probability distribution functions (p.d.f.) from
weighted MC events in which the non-weighted background is also included.
From this p.d.f. a likelihood function for the data can be constructed and a
mass obtained from its maximization. The re-weighting can be applied to
the two di-jet masses independently, thus properly accounting for the event
by event correlations. This gives an improvement of the statistical error of
about 10% with respect to one mass alone[44]'[55].
Shown in Fig. 33 are the effective masses of different final states from
the four LEP experiments, and their comparison with the corresponding
Monte Carlo.
80.6
- LEP1. SLD . vN Data
.... LEP2. pp Data
80.5 68% CL
........., '.
>Q)
.~.
~ 8004
s:
E
80.3
Preliminary
80.2
130 150 170 190 210
mt [GeV]
Figure 35. Comparison of the direct (dashed curve) and indirect (solid curve) determi-
nation of the W mass, and the SM predicted relation as a function of the Higgs mass for
different Higgs masses (shaded area) [41].
44 E. FERNANDEZ
For the 4-quark final state there are two additional problems, as men-
tioned above, namely:
- Color reconnect ion. In the four-quark final-state it is possible that
the two original color singlets interact strongly and exchange color
before hadronization. The effect has been investigated by comparing
the jet charged multiplicities of qqqq final states with those coming
from semileptonic final states qqlv. The latter should not be affected
by the color-reconnection problem. The data are compatible with small
or no effect. The estimated contribution to the systematic error on the
W mass is 25-70 MeV .
6~~--------~~~~~~
\ Aa(5) _
\ had-
t.>
-t O.02804±O.00065
4
o Excluded Preliminary
10 10 3
Figure 36. The line in the figure is the X2 - X;"in from a fit to all the electroweak data
of LEP and the Tevatron. The central value of the Higgs mass is below the exclusion
limit from direct searches, which is 95.2GeV/c2 at 95% CL (shaded area). The 95% CL
upper level from the fit is 245GeV / c2 .
- Bose-Einstein correlations. At short distances there could be correla-

tions between the low-momentum identical bosonic particles such as
neutral pions, leading to an enhancement of their production. This
would lead to momentum transfer between the decay products of the
two bosons, and hence distort the invariant mass. These correlations
have been indeed established by OPAL in hadronic Z decays and in
decays of the same W [15]. The estimated systematic error on Mw is
20-60 MeV.
The error on the beam energy enters directly into the the error of the
W mass, through the kinematic fits:
At present the beam energy is determined by the resonant depolarization

technique at the Z, and then extrapolated to the higher energies. The error
at vis = 183 Ge V is f:lEb = ±50MeV. This error has already a substantial
impact on the final precision.
The results of the four LEP experiment and their averages are shown in
Fig. 34 for the 4-quark and semileptonic final states. The combined result
from LEP is
mw = 80.350 ± 0.056 GeV
It should be compared with those obtained at the Tevatron, by CDF (Mw =
80.433 ± 0.079Gev) and DO (Mw = 80.474 ± 0.093) [15],[39].
Fig. 35 shows the 68% contour of the direct measurements of Mw
versus mt (Tevatron and LEP-2, dashed contour) and the indirect deter-
mination (from LEP-1+SLD+lIN, solid curve). The shaded band is the
SM relationship between Mw and mt. The 3 lines inside correspond to
mH = 95,300, 1000 GeVjc 2 as shown. Again the data favor a small Higgs
mass. In fact, a fit to all the electro-weak precision data from LEP, SLD
and the Tevatron (Fig. 36), gives
mH < 245 GeVjc2 at 95% CL
However, the dependence on the Higgs mass is only logarithmic, and the
upper bound moves up very quickly if we go above 2 standard deviation
limits.
9. Searching for the Standard Model Higgs Boson

We have seen in the above lecture how well the Standard Model works,
explaining consistently the data obtained at LEP and elsewhere. But it is
a fact that the SU(2)L ® U(l)y symmetry of the electro-weak theory is
broken: the photon is massless while the Z and W bosons are very massive,
and all the fermions have mass (except, perhaps, the neutrinos). Under-
standing the mechanism for this breaking, the origin of mass, remains as
one of the fundamental problems in particle physics. The simplest breaking
mechanism is that of the Standard Model itself: a complex iso-doublet field
which spontaneously breaks the symmetry by acquiring a non-zero expec-
tation value. The masses of the particles are generated by their Yukawa-like
interactions with the field. Three of the four components are "used" to cre-
ate the longitudinal components of the Ws and the Z, while the remaining
46 E. FERNANDEZ
fourth manifests itself as a neutral particle: the Higgs. This is a Minimal

Standard Model, and it is very appealing esthetically.
The Yukawa-like interactions implies that the Higgs couples to fermions
proportionally to their masses, which is one of the fundamental properties
that has to be proven if the Higgs is found. Although the mass of the Higgs
is not given by the model, it can be severely constrained by theoretical
considerations (see for example references [2], [56], [57], [58]) .
Consistency conditions within the SM restrict the mass of the Higgs
to values below 1 TeV, while a lower bound on the mass follows from the
requirement that the vacuum be stable. The lower bound depends on the
top mass and on the scale A up to which the SM can be extended before
the emergence of new strong interactions of the fundamental fermions. An
upper bound also follows from the requirement that the SM be valid up to
the scale A,
87l'v 2
m2
H
<
- 3l A2
v:::::: 246GeVjc2 .
og-;r
If we take A up to the Plank scale the limits are
130GeVjc2 < mH < 190GeVjc2

but these limits change if we move A.
Precision measurements of electro-weak observables (particularly at LEP
1) also constrain the possible values of the SM Higgs mass as we have seen in
the previous lecture (see Fig. 36), but given the logarithmic dependence the
resulting values of the Higg mass are not very restrictive. Direct searches
exclude the 8M Higgs with masses below 95.2 GeV jc2 as we will see.
The dominant process for the production of the 8M Higgs at LEP-2
is the "Higgs-strahlung" process, Fig. 37, where the Higgs couples to the
highest mass particle available, the Z. However, for the highest energies the
fusion processes, also depicted in Fig. 37, start to contribute. When
the Higgsstrahlung cross-section drops sharply. This can be seen in Fig. 38,
which shows, on the left, the cross section for Higgs production as a function
of the Higgs mass for three different LEP c.m. energies. This is different
from what happens in hadron machines, like the LHC, where the collisions,
at fixed beam energies, involve a very wide range of energies at the parton
level. Let's take a numerical example: at Vs = 198 GeV and mH =100
Ge V j c2 , the HZ cross-section is about 0.25 pb. If the integrated luminosity
reaches 200pb- 1 , each LEP experiment will collect 50 events. However, one
has to reject the Z Z events and other background, which is important when
e i') e v or e
Z" ) w- or Z
---------H
,,
,, w' or Z
,
e+ H e+ V () r e +
Figure 37. Feynman diagrams for the Higgs-strahlung and fusion (ZZ and WW) pro-
cesses.
the ZZ channel opens, Fig. 38. This translates in an inefficiency of selecting

HZ candidates.
Z 5
.:7 10 2
c
.s:
tl 4
~
10
.-
~
~
--
u .0
3 a.
Q)
Q)
2 b
10- I
14~0~~1~6~0-L~1~8~0--~2~00----2~20--~240
50 60 70 80 90
ffiH (GeV/c 2 ) 'IS (GeV)
Figure 38. On the left the cross-section for Higgs production as a function of the Higgs
mass, for three LEP center of mass energies. On the right the cross-section for several
processes as a function of the center of mass energy.
For Higgs masses in the region of interest for LEP2 (e.g. 90 to 110
GeV jc2 ) the Higgs decays mainly into bb (85%) and much less so into 7+7-
(8%) and cc (4%), and these branching ratios are almost independent of
48 E. FERNANDEZ
the mass, for masses around 100 GeV. The H width is small (e.g. less than
3 MeV for mH less than 100 GeV /c 2). The search strategies are based on
the characteristics of these decays, together with those of a real Z.
All the LEP experiments have searched for the Standard Model Higgs
boson at LEP-2 [59]-[69] (and at LEP-1). There are four different decay
channels to look for which are explained in the next four subsections.
9.1. THE MUON AND ELECTRON CHANNEL
The final state topologies to look for are those of are
Z -t l+l- (l = e, It)
H -t jet - jet
This channel comprises 6.7% of all the Higgs final states (the e+e- channel
has a small contribution from ZZ fusion).
The main selection criteria for this channel are:
- The lepton invariant mass should be close to the Z mass.
- The invariant mass of the system recoiling against the leptons should
be large. (r)
Therefore the events are first selected by requiring identified or isolated
lepton candidates with an invariant mass close to the Z. Background from
WW -t qqlll events are rejected by requiring that the mass of the qq
system be above the W mass. In this channel no b-tag is necessary, and the
efficiency is large ( about 75%).
9.2. THE MISSING ENERGY CHANNEL
Here one looks for final states coming from
Z -t lIii
H -t bb
and this channel represents 17% of the Higgs final states.

The main background for this channel are radiative-return events. The
selection criteria are designed to first reject these events, for example re-
quiring that the missing transverse momentum be in the direction of the
detector (and not along the beam).
The tagging of the jets as coming from b particles is the other very
powerful selection tool (see below) . The b jets from these events are also
acoplanar, as opposed to those coming from radiative Z events which are
planar with the beam.
9.3. THE TAU-JETS TOPOLOGY
The final states are

Z ~ T+T-
H -t bb
or
Z -t qij
H -t T+T-
These events should have 4 jets, 2 of them of low multiplicity. There

is also missing energy due to the neutrinos coming from the decays of the
taus. The first step is to identify the tau jets. Once this is done, the b
tagging is imposed to the other two jets for the first case, T+T- H. For the
H -t 7'+7'- and Z( -t qij) channel, no b-tag is possible and the selection is
less efficient as more stringent cuts are adopted to select the taus.
The total fraction of Higgs events with this topology is about 9% and
selection efficiencies are of the order of 20% for 7'+7'- Hand 17% for the
qijH events.
Background suppression using b-tagging
,
10
., !
DELPHI
~ u ~ M U ~ U U u
Efficiency e+ e - Io hZ
Figure 39. Background rejection power of the b-tagging technique in DELPHI. The
"efficiency for selecting three background processes is shown versus the efficiency for
selecting real Higgs decays.
9.4. THE 4-JET CHANNEL
This channel amounts to 64% of the cases, and consists on

Z -t qij
H -t bb
50 E. FERNANDEZ
The main distinctive characteristic of these events is the 4-jet topology.

The Z and H are produced almost at rest and hence the events should be
"spherical". The jet isolation is one of the main discriminating variables.
(r)
Typical backgrounds are e+e- -+ W+W-, e+e- -+ ZZ, e+e- -+ qij,
and e+ e- -+ qijg where the gluon 9 hadronises into a jet.
The bb requirement suppresses strongly most of the above backgrounds.
The bb tagging technique has been extensively studied at LEP-l (with the
aim of measuring Rb) and its importance is illustrated in Fig. 39.
10. Limits on Higgs mass

The latest data from the LEP experiments presented at the Tampere Con-
ference [70],[71] are shown in Table. 3. No significant number of events were
found above what is expected from background processes.
Luminosity Predicted Background Events Observed
ALEPH 176.2 44.4 53

DELPHI 158 172.7 187
L3 176.4 91.1 94
OPAL 172.1 35.4 41
TABLE 3. Higgs candidates expected from background processes and

candidates observed in the data on the four LEP experiments [70], [71].
From the numbers on the table there remain the problem of giving a
lower bound on the mass of the Higgs, combining the four LEP experiments.
The issue is complicated since, on the one hand, the expected signal is
well below background, and, on the other, the individual limits from the
experiments, or from a single experiment but obtained with several different
methods, have to be combined in a statistically consistent way. There has
been a lively debate on how this can be done in an optimal way and the
reader is referred to references [72), [73]. Again an LEP-wide group was
formed to combine the results of the four experiments (see [74]) . The limit
is obtained incorporating not only the assumed fluctuations of the number
of background events, but also their characteristics (e.g. effective masses of
the candidate Higgs decay products).
The limit presented at Tampere is
mH < 95.2GeVjc2
Both points should be shifted

200 Me V to the left.
expected.
500 observed *
400 ALEPH
~
V)
300
DELPHI
o
s::: 200 L3
E OPAL
:J 100
....J
50
80 85 90 95 100 105 110

MH (Gev)
Figure 40. The luminosity required to establish a Standard Model Higgs signal at the
5a level as a function of the Higgs mass, for the indicated center of mass energies [70].
The limit is significantly above the Z mass, that is a Higgs with a mass
smaller or equal than that of the Z is strongly excluded. The limit above is
actually the "observed" limit, while the "expected limit" (what one would
expect assuming that there is no signal and that all we have is background
with appropriate fluctuations) is 97.2 Ge V / c2 . An "observed" limit smaller
than the "expected" , means that the data do have characteristics of a signal,
that is, that the hypothesis that they are all background is not as good as
it should, statiscally, be!.
The above is indeed the remaining question: what is the minimal mass
that the Higgs should have such that it could still be found at LEP 2? The
answer depends of course on the maximum energy that LEP can reach and
on the integrated luminosity that it can deliver. This has been investigated,
for example in [73J. The conclusion is that for an integrated luminosity
of 200 pb- 1 at a center of mass energy of 200 GeV, the exclusion limit
(the minimum mass of the Higgs such that the signal hypothesis can be
excluded) is 109.1 GeV /c 2 , while the discovery potential (the maximum
mass of the Higgs such that the signal hypothesis can be established at
the 50" level) is 106.9 GeV /c 2 . This is ilustrated in Fig. 40 which shows
52 E. FERNANDEZ
the luminosity required to establish a Standard Model Higgs signal at the

5a level as a function of the Higgs mass, for the indicated center of mass
energies [70].
11. Acknowledgments
I thank Jean-Jacques Aubert for the invitation to give the lectures and for
his hospitality in Cargese. My warm thanks also to Raymond Gastmans and
Jean-Mark Gerard for the very pleasant running of the school at such beau-
tiful location. I also thank Eilam Gross, Alex Read and Eusebio Sanchez
for the copy of the transparencies of their presentations at Tampere.
12. References
1. Physics at LEP, eds. J. Ellis and R. Peccei, CERN 86-02 (1986).
2. Physics at LEP2, eds. G. Altarelli, T. Sjostrand and F. Zwirner, CERN
96-01 (1996).
3. ALEPH Collaboration, Phys. Lett. B378 (1996) 373
5. ALEPH Collaboration, CERN-EP-99-042
6. DELPHI Collaboration, CERN-EP-99-005
7. DELPHI Collaboration, Phys. Lett. B456(1999) 310
8. L3 Collaboration, Phys. Lett. B370 (1996) 195
11. OPAL Collaboration, Phys. Lett. B391 (1997) 221
12. OPAL Collaboration, Eur. Phys. J. C2 (1998) 441
14. OPAL Collaboration, CERN-EP-99-097
15. J. Mnich Plenary talk on "Tests of the Standard Model" at the EPS-
HEP Conference in Tampere, Finland, July 1999.
19. DELPHI Collaboration, Phys. Lett. B397 (1997) 158
20. DELPHI Collaboration, Eur . Phys. J. C2 (1998) 581
21. Phys. Lett. B398 (1997)223
22. Phys. Lett. B407 (1997)419
23. Phys. Lett. B436 (1998)437
27. ALEPH Collaboration, CERN-EP-98-178

37. M. Campanelli, Physics of W bosons at LEP2, Int. J. Mod. Phys. A14
(1999) 3277
38. Talk of E. Sanchez at the EPS-HEP Conference in Tampere, Finland,
July 1999.
39. See for example the review of H. Montgomery in these proceedings.
40. M. Martinez, R. Miquel, L. Rolandi and R. Tenchini: "Precision Tests
of the Electroweak Interaction at the Z Pole", Review of Modern
Physics, 71 (1999) 575.
41. The LEP electroweak working group CERN-EP /99-15, and updates
for the summer conferences.)
46. DELPHI Collaboration, Eur. Phys. J. C2 (1998) 581
51. OPAL Collaboration, Eur. Phys. J. Cl (1998) 395
53. T. Muta, R. Najima ans S. Wakaizumi, Mod. Phys. Lett. Al (1986)
203.
54. A. Juste, Measurement of the W mass in e+e- annihilation, Thesis,
Univ. Autonoma de Barcelona, 1998 (unpublished).
55. 1. Riu, Measurement of the W mass from the WW -+ qqqq channel
with the ALEPH Detector, Thesis, Univ. Autonoma de Barcelona,
1998 (unpublished).
56. M. Spira and P. Zerwas, "Electroweak Symmetry Breaking and Higgs
Physics" CERT-TH/97-379,DESY 97-261,hep-ph/9803257.
57. J. Ellis, "Phenomenology of LEP2 Physics", Lecture at Lake Louise
Winter Institute, Feb. 1977, CERN-TH/97-131.
58. F. Richard, "Search for Higgs Bosons in e+ e- Colliders. LAL 98-74.
54 E. FERNANDEZ
59. ALEPH Collaboration, Phys. Lett. B412 (1997) 155.

62. DELPHI Collaboration, Eur. Phys. J. C2 (1998) 1.
63. DELPHI Collaboration, Eur. Phys. J. CI0 (1999) 563.
64. L3 Collaboration, Phys. Lett. B411 (1997) 373.
67. OPAL Collaboration, Phys. Lett. B393 (1997) 231.
68. OPAL Collaboration, Eur. Phys. J. Cl (1998) 425.
69. OPAL Collaboration, Eur. Phys. J. C7 (1999) 407.
70. Talk of A. Read at the EPS-HEP Conference in Tampere, Finland,
July 1999.
71. E. Gross Plenary talk on "Searches for New Particles" at the EPS-HEP
Conference in Tampere, Finland, July 1999.
72. P. Janot and F. Le Diberder, "Optimally Combined Confidence Lim-
its", Nucl. Instrum. and Methods A411 (1998) 449.
73. E. Gross, A. L. Read and D. Lellouch, "Prospects for the Higgs Boson
Search in e+e- Collisions at LEP 200", CERN-EP /98-094.
74. LEP working group for Higgs searches, CERN-EP /99-060.
ANOMALOUS QUARTIC COUPLINGS AT LEP2
W. JAMES STIRLING
Departments of Physics and Mathematical Sciences
University of Durham
Durham DHl 3LE, u.K.
AND
ANJA WERTHENBACH
Department of Physics
Durham DHl 3LE, U.K.
Abstract. The production of three electroweak gauge bosons in high-

energy e+ e- collisions offers a window on anomalous quartic gauge bo-
son couplings. We investigate the effect of three possible anomalous cou-
plings on the cross sections for W+W-, and ZO" productions at LEP2
(y's = 200 GeV).
1. Introduction
In the Standard Model (SM), the couplings of the gauge bosons and fermions
are tightly constrained by the requirements of gauge symmetry. In the
electroweak sector, for example, this leads to trilinear VVV and quartic
VVVV interactions between the gauge bosons V = "Zo, W± with com-
pletely specified couplings. Electroweak symmetry breaking via the Higgs
mechanism gives rise to additional Higgs - gauge boson interactions, again
with specified couplings.
The trilinear and quartic gauge boson couplings probe different aspects
of the weak interactions. The trilinear couplings directly test the non-
Abelian gauge structure, and possible deviations from the SM forms have
been extensively studied in the literature, see for example [1] and references
55
56 W. JAMES STIRLING AND ANJA WERTHENBACH
therein. Experimental bounds have also been obtained [2]. In contrast, the
quartic couplings can be regarded as a more direct window on electroweak
symmetry breaking, in particular to the scalar sector of the theory (see
for example [3]) or, more generally, on new physics which couples to elec-
troweak bosons.
In this respect it is quite possible that the quartic couplings deviate

from their SM values while the triple gauge vertices do not. For exam-
ple, if the mechanism for electroweak symmetry breaking does not reveal
itself through the discovery of new particles such as the Higgs boson, su-
persymmetric particles or technipions it is possible that anomalous quartic
couplings could provide the first evidence of new physics in this sector of
the electroweak theory [3J.
High-energy colliders provide the natural environment for studying anom-

alous quartic couplings. The paradigm process is f J --+ VVV, with f = e
(e+e- colliders) or f = q (hadron-hadron colliders), where one of the Feyn-
man diagrams corresponds to f J --+ V* --+ VVV. In this context, one may
consider the quartic-coupling diagram(s) as the signal, while the remaining
diagrams constituting the background. The sensitivity of a given process to
anomalous quartic couplings depends on the relative importance of these
two types of contribution, as we shall see.
In this study we shall focus on e+ e- collisions, and quantify the depen-

dence of various e+ e- --+ VVV cross sections on the anomalous couplings.
We shall consider in particular v's = 200 GeV, corresponding to LEP2. For
obvious kinematic reasons, processes where at least one of the gauge bosons
is a photon have the largest cross sections. Indeed, VVV production with
V = Zo, W± are kinematically forbidden at 200 GeV. Also ZO Z°'Y pro-
duction is highly suppressed at 200 GeV. We therefore consider W+W-'Y
and Z°'Y'Y production. Each of these contains at least one type of quartic
interaction. 1
There have been several studies of this type reported in the literature
[4, 5]. Our aim is partly to complete as well as update these, and partly
to assess the relative merits of the above-mentioned processes in providing
information on the anomalous couplings. Note that our primary interest is
in the so-called 'genuine' anomalous quartic couplings, i.e. those which give
no contribution to the trilinear vertices.
1 We ignore the process e+e - -+ ''(''I, which involves no trilinear or quartic interactions.
ANOMALOUS QUARTIC COUPLINGS AT LEP2 57
In the following section we review the various types of anomalous quar-

tic coupling that might be expected in extensions of the SM. In Section 3
we present numerical studies illustrating the impact of the anomalous cou-
plings on various VVV cross sections. Finally in Section 4 we present our
conclusions.
2. Anomalous gauge boson couplings

The lowest dimension operators which lead to genuine quartic couplings
where at least one photon is involved are of dimension 6 [4J . A dimension
4 operator is not realised since a custodial SU(2) symmetry is required to
keep the p parameter, p = Mar / (M1 cos 2 Ow) , close to its measured SM
value of 1. Thus the 4-dimensional operator
1 --t --t --t --t
£4 = -4 g2(Wtt x W v ) (wtt x WV) (1)
with
~(W:+W;)
--t
W tt =
( Jz(W: - W;)
W tt3 - fiB
9 tt
) (2)
and
I
W3-~B
tt 9 tt
= cos(}w Ztt + sin (}wAtt - ~() sin(}w (- sin (}wZtt + cos (}wAtt)
cos w e
= ~ (3)
cos Ow
does not involve the photon field Aw The other possible 4-dimensional
operator [4J
r --
~4
. M2
-ze ..\)' FttVWtttQ W Q
v (4)
W
with
Ftt V = 0ttAv - ovAtt
W ttV 0ttWV - OvWtt - gW tt x Wv (5)
and
(6)
generates trilinear couplings in addition to quartic ones and is therefore

not 'genuine'. In Section 4 we will briefly discuss the impact of possible
non-zero anomalous trilinear couplings on our analysis.
We are therefore left with several 6-dimensional operators. First the

neutral and the charged Lagrangians, both giving anomalous contributions
to the VV TY vertex, with VV either being W+W- or ZO Zo.
e2 --+ --+
LO = -16A2 ao FI)'v FJ.Lv W a . Wa
e2
-16A2 ao [ - 2(PI . P2)(A . A) + 2(PI . A)(p2 . A)]
x[2(W+ . W-) + (Z . Z)/ cos 2 Ow] (7)
e2 --+ ---+
Lc = -- - a FJ.La FJ.Lv
16A2 c
I'l W(3 . W.a
2
- 1:A2 a c [ - (PI' P2) All A(3 + (PI' A) Aap2(3
+ (P2 . A) pf A(3 - (A· A) pfP2(3]
x [W;W+(3 + W';W-(3 + ZaZ(3jcos2 Ow] . (8)
where PI and P2 are the photon momenta.
Since we are interested in the anomalous VV" contribution we pick

up the corresponding part of the Lagrangian. To obtain the Feynman rules
for the corresponding vertex (in agreement with [6]) we have to multiply
by 2 for the two identical photons (as well as for the ZOs in the case of
VV = ZO Zo) and by i for convention.
Finally, an anomalous WW Z, vertex is obtained from the Lagrangian

2
L
n
= i_e_ a E" W(i)W(j)W(k)a FJ.LV
16A2 n tJk J.La v
e2
= a (pV AJ.L _ pJ.L AV)
16A2 cos Ow n
x (-W;p~ (Z· W+) + W:P~ (Z· W-) + ZvP~ (W+ . W-)
-ZvP~ (W+ . W-) + W;W: (p+ . Z) - W:W; (p- . Z)
-ZvW: (p+ . W-) + ZvW; (p- . W+) - W:p~ (Z . W-)
+W;p~ (Z . W+) - W; ZJ.L (pO. W+) + W: ZJ.L (pO. W-)) (9)
where wY) are the components of the vector (2) and p, p+, p- and pO are
the momenta of the photon, the W+, the W- and the ZO respectively.
It follows from the Feynman rules that any anomalous contribution is

linear in the photon energy E"(. This means that it is the hard tail of the
photon energy distribution that is most affected by the anomalous contri-
butions, but unfortunately the cross sections here are very small. In the
following numerical studies we will impose a lower energy photon cut of
E~in = 20 GeV. Similarly, there is also no anomalous contribution to the
initial state photon radiation, and so the effects are largest for centrally-
produced photons. We therefore impose an additional cut of Irhl < 2.2
A further consideration concerns the effects of beam polarisation. One

of the 'background' (Le. non-anomalous) diagrams for e+e- --t W+W-,
production is where all three gauge bosons are attached to the electron line.
Such contributions can be suppressed by an appropriate choice of beam po-
larisation (Le. right-handed electrons) thus enhancing the anomalous signal.
Unfortunately this is not possible at LEP2.
Finally, the anomalous parameter A that appears in all the above anoma-
lous contributions has to be fixed. In practice, A can only be meaning-
fully specified in the context of a specific model for the new physics giv-
ing rise to the quartic couplings. One example is an excited W scenario
W+, --t W* --t W+" where we would expect A rv Mw* and ai to be
related to the decay width for W* --t W +,. However, in order to make our
analysis independent of any such model, we choose to fix A at a reference
value of Mw, following the conventions adopted in the literature. Any other
choice of A (e.g. A = 1 Te V) results in a trivial rescaling of the anomalous
parameters ao, a c and an·
3. Numerical studies
In this section we study the dependence of the cross sections on the three
anomalous couplings defined in Section 2. As already stated, we apply a cut
on the photon energy E"( > 20 GeV to take care of the infrared singularity,
and a cut on the photon rapidity Irhl < 2 to avoid collinear singularities.
We do not include any branching ratios or acceptance cuts on the decay
products of the produced W± and ZO bosons, since we assume that at e+ e-
colliders the efficiency for detecting these is high.
20bviously in practice these cuts will also be tuned to the detector capabilities.
Next we study the influence of each of the three anomalous parame-

ters ao, ac and an separately in order to gauge the impact of each on the
cross section. Note that a(W+W-'Y) depends on all three parameters, while
a(Z°'Y'Y) depends only on ao and ac . The dependence on the ai is quadratic,
since they appear linearly in the matrix element; the interference between
the anomalous and standard parts of the matrix element is small. The
anomalous parameters have a markedly different effect on the three cross
sections. The anomalous process e+ e- -+ ZO* -+ Z°'Y'Y has a much larger
impact on a(Z°'Y'Y) since there are only six other SM diagrams. In contrast,
e+e- -+ 'Y* -+ W+W-'Y has a much larger SM 'background' set of diagrams
to contend with. Note also that the anomalous contributions are enhanced
by a factor 1/ cos 4 Ow compared to the WW'Y'Y vertex.
Of course the important question is which of the two processes offers

the best chance of detecting an anomalous quartic coupling. To answer
this we need to combine information to see whether enhanced sensitivity
can overcome a smaller overall event rate. We also need to consider corre-
lations between different anomalous contributions to the same cross section.
We consider the experimental scenario of unpolarised e+ e- collisions at

200 GeV with J £ = 150 pb- I . Starting with the W+W-'Y process, Fig-
ure 1 shows the contours in the (ai,aj) plane that correspond to +2,+3a
deviations from the SM cross section at ..;s= 200 GeV. Note that there
are three ellipses, one for each combination of the three anomalous cou-
plings. Evidently the sensitivity to ao and an is comparable, corresponding
to ai < 0(100) for this luminosity. The corresponding limit on ac is some
three to four times larger.
Turning to the sensitivity of the Z°'Y'Y process, Figure 2 shows the sen-
sitivity of Z°'Y'Y to ao and ac at ..;s = 200 GeV with J £ = 150 pb- I .
For comparison, we also show the corresponding W+W-'Y contours from
Figure 1. The Z°'Y'Y process gives a significant improvement in sensitivity,
particularly for a c . Since the SM cross sections at this energy are compa-
rable the improvement comes mainly from the enhanced sensitivity of the
matrix element to the anomalous couplings in the Z°'Y'Y case.
4. Discussion and Conclusions
We have investigated the sensitivity of the processes e+ e- -+ W+W- 'Y and

Z°'Y'Y to genuine anomalous quartic couplings (ao, ac , an) at the canonical
centre-of-mass energy ..;s = 200 GeV (LEP2). Key features in determin-
ing the sensitivity for a given process, apart from the fundamental process
dynamics, are the available photon energy E" the ratio of anomalous dia-
grams to 8M 'background' diagrams, and the polarisation state of the weak
bosons [4].
The process e+ e- -7 ZO" leads to the tightest bounds on the contour

of (ao, ac ), while the process e+ e- -7 W+W-, is needed to set bounds also
on an' Note that the contours of (ao, an) and (a c , an) can then be improved
using the know ledge of the tighter bounds on the contour of (ao , ac ) from
.... i=O. j=c

1000 ••,. , , , ... ,J,. • .....
i=O. j = n
"
•••• ~-'
-, ..- / .....
.' •• 1't,,:' , - • -, '" -', ••••
i=c. j =n
.. -. . ..
\
.... . ~,,,
/.. ?
/..- /
'I '\\
,, ,,
, ,
\
J ,
.....
,
500
----
--
: , j
.,
:" I
: I I
:' :-'
: ,
SM
i5 0
,
..,. .,,
, ,
,, , ..
, ,
- 500 "
..... I. '. , . .,
,, ,,
,
', I ,
,, ,
-" " \ ,, ,,
'>-. \
1
\
',, \
-..~ ... ." ,
- 1000 .... ...., , ",' ..... -..: ..... -:.' .....•.•.'... ,- .'
oute r: 99 .73% CL
"'~ """'';''':'-''''''', ~,~~ , in ner : 95.45 % CL
- 300 -200 - 100 o 100 200 300 400 500

OJ
Figure 1. Contour plots for +2,+3(1 deviations from the SM e+e- -7 W+W-, total
cross section at Vs = 200 GeV with £ = 150 pb- 1 , when two of the three anomalous J
couplings are non-zero.
500
400
e+e- w'w--y
e+e- -7 z-yy
300
200
u
o 100
-100
outer : 99.73 % CL
inner : 95.45 % CL
-200
- 200 - 100 o 100 200 300 400 soc

00
Figure 2. Contour plots for +2,+30" deviations from the 8M e+e- -+ Z°-y-y total cross
f
section at Vs = 200 GeV with C = 150 pb-l. For comparison, the corresponding
contours for the e+e- -+ W+W--y process from Figure 1 are also shown.
ZO" production. At this energy ZO" benefits kinematically from produc-

ing only one massive boson, which leaves more energy for the photons as
well as having fewer 'background' diagrams. On the other hand W+W-,
production at this energy suffers from the lack of phase space available for
energetic photon emission, although this is partially compensated by the
production of longitudinal bosons, which gives rise to higher sensitivity to
the anomalous couplings.
Finally it is important to emphasise that in our study we have only

considered 'genuine' quartic couplings from new six-dimensional operators.
We have assumed that all other anomalous couplings are zero, including the
trilinear ones. Since the number of possible couplings and correlations is so
large, it is in practice very difficult to do a combined analysis of all cou-
plings simultaneously. In fact, it is not too difficult to think of new physics
scenarios in which effects are only manifest in the quartic interactions. One
example would be a very heavy excited W resonance produced and decay-
ing as in W+')' -t W* -t W+')'.
In principle, any non-zero trilinear coupling could affect the limits ob-
tained on the quartic couplings. For example, in equation (4) we showed
explicitly how a non-zero trilinear coupling (A) can generate an anomalous
WW,),,), quartic interaction to compete with the 'genuine' ones that we have
considered. The (dimensionless) strength of the former is egA, while for the
latter it is e2ai(Eext.)(Eind/A2, where E ext . and Eint. are the typical en-
ergy scales of the photons entering the vertex. (Here we are considering,
as a specific example, the e+e- -t W+W-,), process.) Since A = Mw,
(Eext.) '" 25 GeV and Eint. '" [5yS + 4{ yS - (E ext .) )]/9 '" 190 GeV , both
for yS = 200 GeV, we see immediately that the relative contributions of the
two types of couplings are in the approximate ratio 3A : ai. Now, at LEP2
upper limits on trilinear couplings like A are already 0{0.1) [2]. In contrast,
we have shown that the limits achievable on the ai are 0(100). Hence we
already know that the anomalous trilinear couplings have a minimal impact
on our analysis. The same argument holds at higher collider energies. The
limits on the trilinear couplings will always be so much smaller than those
on the quartic couplings, that they can safely be ignored in studies of the
latter.
Acknowledgements
This work was supported in part by the ED Fourth Framework Programme
'Training and Mobility of Researchers', Network 'Quantum Chromodynam-
ics and the Deep Structure of Elementary Particles', contract FMRX-CT98-
0194 (DG 12 - MIHT). AW gratefully acknowledges financial support in the
form of a 'DAAD Doktorandenstipendium im Rahmen des gemeinsamen
Hochschulprogramms III fur Bund und Lander'.
References
1. K. Hagiwara, R.D. Peccei, D. Zeppenfeld and K. Hikasa, Nucl. Phys. B282 (1987)
253.
Triple Gauge Boson Couplings, G. Gounaris et al., in 'Physics at LEP2', Vol. 1,
p. 525-576, CERN (1995) [hep-ph/9601233] .
2. ALEPH Collaboration: R. Barate et al., Phys. Lett. B422 (1998) 369; pre print
CERN-EP-98-178, November 1998 [hep-ex/9901030J.
OPAL Collaboration: G. Abbiendi et al., preprint CERN-EP-98-167, October 1998
[hep-ex/9811028J.
3. S. Godfrey, Quartic Gauge Boson Couplings, Proc. International Symposium on
Vector Boson Self-Interactions, UCLA, February 1995.
4. G. Belanger and F. Boudjema, Phys. Lett. B288 (1992) 201.
5. G. Abu-Leil and W.J. Stirling, J. Phys. G21 (1995) 517.
6. O.J.P. Eboli, M.C. Gonzalez-Gracia and S.F. Novaes, Nucl. Phys. B411 (1994) 381.
7. V. Barger et al., Phys. Rev. D39 (1989) 146.
8. T . Stelzer and W.F. Long, Comput . Phys. Commun. 81 (1994) 357.
9. B.H. Wiik, The TESLA Project, Lecture at the Ettore Majorana school 'From the
Planck Length to the Hubble Radius', Erice, Sept. 98, to appear in the proceedings.
PHYSICS WITH AN e+ e- LINEAR COLLIDER
AT HIGH LUMINOSITY
P.M. ZERWAS
Deutsches Elektronen- Synchrotron DESY
D-22603 Hamburg, Germany
Abstract. The physics potential is summarized briefly for an e+e- linear

collider operating at center-of-mass energies up to Vs = 1 Te V and at inte-
grated luminosities up to J .c = 0.5 ab- 1 per year. This machine will allow
to perform precision studies of the top quark and the electroweak gauge
bosons at the per-mille level. It will be an ideal instrument to investigate
the properties of the Higgs boson and to establish essential elements of
the Higgs mechanism sui generis. In the area beyond the Standard Model,
new particles and their interactions can be explored comprehensively. In
supersymmetric theories, the mechanism of the symmetry breaking can be
investigated experimentally and the underlying unified theory can be re-
constructed eventually at a scale near the Planck scale. Such an e+ e- linear
collider provides unique opportunities in areas complementary to the LHC,
and the high precision allows stable extrapolations up to scales near the
Planck scale.
1. Synopsis
The knowledge of the fundamental constituents of matter and their inter-

actions has been greatly advanced in the past three decades by operating
e+ e- colliders. A coherent picture of the structure of matter has emerged,
that is adequately described by the Standard Model, in many of its facets
at a level of unprecedented accuracy. However, the Standard Model does
not provide us with a comprehensive theory of matter. Neither the funda-
mental parameters, masses and couplings, nor the symmetry pattern can
be explained, but they are merely built into the model. Moreover, gravity
is not incorporated at the quantum level. First steps to solutions of these
problems are associated with the unification of the electroweak and the
65
1.-1. Aubert et a/. (eds.), Particle Physics: Ideas and Recent Developments, 65-93.
66 P.M. ZERWAS
strong forces, and with the supersymmetric extension of the model which
provides a bridge from the presently explored energy scales up to scales
close to the Planck mass.
Two strategies can be followed to enter into the area beyond the Stan-
dard Model. (i) Properties of the particles and forces within the Standard
Model will be affected by new energy scales. Precision studies of the top
quark, the electroweak gauge bosons and the Higgs boson can thus re-
veal clues to the physics beyond the Standard Model. (ii) Above the mass
thresholds, new phenomena can be searched for directly and studied thor-
oughly so that the underlying basic theories can be reconstructed.
In this dual picture a variety of fundamental problems still remain to

be solved within the Standard Model [1, 2], demanding experiments at en-
ergies beyond the range of existing accelerators.
(i) The mass of the top quark is much larger than the masses of the elec-
troweak gauge bosons. Understanding the role of this particle in Nature is
therefore an important goal for the future. In the tl threshold region of e+ e-
collisions the top quark mass can be measured to an accuracy better than
200 MeV. This is a desirable goal since future theories of flavor dynamics
should provide relations among the lepton/quark masses and mixing angles
in which the heavy top quark is expected to playa key role. In addition,
stringent tests in the electroweak and Higgs sector of the Standard Model
can be carried out when the top mass is known very accurately. Helicity
analyses of the tl production vertex and the t decay vertex will determine
the magnetic dipole moments of the top quark and the chirality of the (tb)
decay current at the percent level. Bounds on the CP violating electric
dipole moments of the t quark can be set to 10- 18 ecm.
(ii) Studying the dynamics of the electroweak gauge bosons is another im-
portant task at high energy e+ e- colliders. The form and the strength of
the triple and quartic couplings of these particles are uniquely predicted
by the non-abelian gauge symmetry of the theory, defining the electroweak
charges, the magnetic dipole moments and the electric quadrupole moments
of the W± bosons in the static limit. Tests of the fundamental symmetry
concepts can be performed at an accuracy of 10- 3 down to 10- 4 .
(iii) A high- luminosity e+ e- collider with an energy between 300 and
500 Ge V will be an ideal instrument to search for Higgs particles [3] through-
out the mass range characterized by the scale of electroweak symmetry
breaking, and to investigate their properties. The intermediate Higgs mass
range below ""' 200 GeV is the theoretically and experimentally preferred
region. In this scenario Higgs particles remain weakly interacting up to the
scale of grand unification, thus providing a basis for the renormalization of
the electroweak mixing angle sin2 Ow from the GUT symmetry value 3/8
PHYSICS WITH AN e+e- LINEAR COLLIDER 67
down to the experimentally observed value close to 0.2. Once the Higgs par-
ticle is found, its properties can be studied thoroughly, the external quan-
tum numbers ,fpc and the Higgs couplings, including the self- couplings of
the particle. The measurements of these couplings are the necessary ingre-
dient to establish the Higgs mechanism sui generis experimentally.
Even though many facets of the Standard Model are experimentally

supported at a level of very high accuracy, extensions should nevertheless
be anticipated as argued before. The next generation of accelerators can
shed light on three domains in the area beyond the Standard Model.
The Grand Unification of the gauge symmetries [4] suggests itself quite
naturally. This idea can be realized in different scenarios some of which
predict new vector bosons and a plethora of new fermions. Mass scales of
these novel particles could be as low as a few hundred Ge V.
A very important theoretical extension of the Standard Model, which is
interrelated with the unification of the gauge symmetries, is Supersymmetry
[5]. This novel symmetry concept unifies matter and forces by pairing the
associated fermionic and bosonic particles in novel multiplets. Several ar-
guments strongly support the hypothesis that this symmetry is realized in
Nature. (i) Supersymmetry stabilizes light masses of Higgs particles in the
context of very high energy scales as demanded by grand unified theories.
(ii) Supersymmetry may generate the Higgs mechanism itself by inducing
the radiative symmetry breaking of SU(2)L x U(l)y while leaving U(l)EM
and SU(3)c unbroken for a top quark mass between 150 and 200 GeV. (iii)
This symmetry picture is also strongly supported by the successful predic-
tion of the electroweak mixing angle in the minimal version of the theory.
The particle spectrum in this theory drives the evolution of the electroweak
mixing angle from the GUT value 3/8 down to sin2 Ow = 0.2336 ± 0.0017
which is within a margin of "-' 0.002 to the experimental value sin2 O~{? =
0.2316 ± 0.0002.
In supersymmetric theories a spectrum of several neutral and charged
Higgs bosons is predicted. The mass of the lightest Higgs boson is in nearly
all scenarios less than "-' 150 GeV while the heavy Higgs particles have
masses of the order of the electroweak symmetry breaking scale. Many other
novel particles are predicted in supersymmetric theories. The scalar part-
ners of the leptons could have masses in the range of "-' 200 GeV whereas
squarks are expected to be considerably heavier. The lightest supersymmet-
ric states are likely to be non-colored gaugino/higgsino states with masses
possibly in the 100 GeV range. Searching for these supersymmetric particles
will be one of the most important tasks at future e+e- colliders. Moreover,
the high accuracy which can be achieved when masses and couplings are
measured, will allow us to determine the mechanism of supersymmetry
68 P.M. ZERWAS
breaking and to extrapolate the basic parameters of the theory so that the
key elements of the underlying grand unified theories at scales potentially
close to the Planck scale can be reconstructed.
While new high-mass vector bosons and particles carrying color quan-
tum numbers can be searched for very efficiently at hadron colliders, e+ e-
colliders provide in many ways unique opportunities to discover and ex-
plore non-colored particles. This is most obvious in supersymmetric theo-
ries. Combining LEP2 analyses with future searches at the Tevatron and
the LHC, the light and heavy Higgs bosons can individually be found only
in part of the supersymmetry parameter space. Squarks and gluinos can
be searched for very efficiently at the LHC. Yet precision studies of their
properties are possible only in part of the parameter space. Similarly non-
colored supersymmetric particles; a model-independent conclusive analysis
of gauginosjhiggsinos and scalar sleptons can only be carried out at e+ e-
colliders with well- defined kinematics at the level of the subprocesses. The
detailed knowledge of all the properties of the colored and non-colored
supersymmetric states will enable us to reveal the mechanism of supersym-
metry breaking and the structure of the underlying theory.
Thus, the physics progamme of e+e- linear colliders is in many aspects
complementary to the programme of the pp collider LHC. The high accu-
racy which can be achieved at e+ e- colliders in exploring the properties
of the top quark, electroweak gauge bosons, Higgs particles and supersym-
metric particles will enable us to cover the energy range above the existing
machines up to the Te V region in a conclusive form, eventually providing
us with essential clues to the basic structure of matter and of the laws of
physics.
The discussion will focus on the physics program at an e+ e- linear col-
lider operating at center-of-mass energies above LEP2 up to about 1 TeV.
Primarily the potential high-luminosity runs, collecting integrated lumi-
nosities up to 0.5 ab- 1 per annum will be described. Also the results ex-
pected from high-luminosity runs at low energies on the Z resonance, the
GigaZ mode, and near the WW threshold will be summarized. Electrons
and positrons will in general be assumed polarized to 80% and 60%, re-
spectively. Specific problems which can be solved in el and II modes of
the linear collider will be addressed in the appropriate context.
This summary report is built on the general linear collider review Ref.[6].
Other material can be found in Refs. [7] , experimental aspects in particular
in Ref.[8J. For recent summaries of LHC physics and J.LJ.L physics see Refs.[9]
and [10]' respectively.
2. Top Quark Physics

Top quarks are the heaviest matter particles in the 3- family Standard
Model. Introduced to incorporate CP violation in the left- handed charged
current sector, they may therefore hold the key for aspects of the physics
beyond the Standard Model at high- energy scales. Examples in which the
large t mass is crucial, are mUlti- Higgs doublet models, models of dynami-
cal symmetry breaking, compositeness and supersymmetry. Strong indirect
evidence for the existence of top quarks, based on the well established
gauge symmetry pattern of the Standard Model, had been accumulated
quite early. By evaluating the high-precision electroweak data, the value of
the top quark mass was estimated to be mt = 180 ± 14 GeV. Top quarks
have recently been observed directly by the two Tevatron experiments [11],
corresponding to a mass of mt = 173.8 ± 5.2 Ge V which is in striking
agreement with the result of the electroweak analysis.
2.1. PROFILE OF THE TOP QUARK
For a top mass larger than the W mass, the channel
t -t b+ W+
is the dominant decay mode. For mt '" 175 GeV the width of the top quark,
rt '" 1.4 Ge V, is so large compared with the scale A of the strong interac-
tions that this quark can be treated as a bare quantum which is not dressed
by non-pertubative strong interactions [12].
Chirality of the (tb) decay current: The precise determination of the weak
isospin quantum numbers does not allow for large deviations of the (tb)
decay current from the left-handed SM prescription. Nevertheless, since
V + A admixtures may grow with the masses of the quarks involved [through
mixing with heavy mirror quarks, for instance], it is necessary to check the
chirality of the decay current directly. The 1+ energy distribution in the
semileptonic decay chain t -t W+ -t 1+ depends on the chirality of the
current. Any deviation from the standard V - A current would lead to a
stiffening of the spectrum and, in particular, to a non-zero value at the
upper end-point of the energy distribution. A sensitivity of about 5% to a
possible V + A admixture can be reached experimentally [13].
Non- standard top decays could occur in supersymmetric extensions of the
Standard Model: top decays into charged Higgs bosons and/or top decays
to stop particles, t -t b + H+ and t -t t + X~. If kinematically allowed,
branching ratios of order 10% are expected in both cases so that these
decay modes could be observed easily. Decays with signatures as clean as
70 P.M. ZERWAS
t ~ Cf, cZ may be detected for branching ratios of order 10- 4 and less.
The main production mechanism for top quarks in e+ e- collisions is the

annihilation channel [14],
e+e - I,Z t-t
~
For mt "-J175 GeV, the maximum of the cross section a(it) 800 fb is
"-J
reached about 30 GeV above the threshold, giving rise to a million top
quarks in two years of collider operation. If the scale of new areas beyond
the Standard Model is much larger than the collider energy, the electroweak
production currents can globally be described by form factors which reduce
to anomalous Z charges, anomalous magnetic dipole moments and electric
dipole moments.
Magnetic dipole moments: If the electrons in e+ e- ~ it are left- handedly
polarized, the top quarks are produced preferentially as left-handed par-
ticles in the forward direction while only a small fraction is produced as
right-handed particles in the backward direction [15], so that the backward
direction is most sensitive to small anomalous magnetic moments of the top
quarks. The anomalous magnetic moments can be bounded to the percent
level by measuring the angular dependence of the t quark cross section in
this region.
Electric dipole moments: These moments are generated by CP-non invari-
ant interactions. Non-zero values of the moments can be detected through
non-vanishing expectation values of CP-odd momentum tensors such as
Tij = (q+ - q-)i(q+ x q_)j with q± being unit momentum vectors of
the W decay leptons. Sensitivity limits to " Z electric dipole moments
of ;S 10- 18 e cm can be reached [16] for an integrated luminosity of J £ =
100 fb- 1 at y's = 500 GeV.
2.2. THE TOP MASS
Quark-antiquark production near the threshold in e+ e- collisions is of

exceptional interest. The long time which the particles stay close together
at low velocities, allows the strong interactions to build up rich structures of
bound states and resonances. This picture would have applied to top quarks
up to the mass range of rv 130 GeV. Beyond this value, the picture changes
quite dramatically as a result of the rapid top decay: The decay time of the
states becomes shorter than the revolution time of the constituents so that
top onium resonances cannot be formed any more [12]. For a while, however,
remnants of the IS state give rise to a peak in the excitation curve, yet
it disappears for t masses in excess of 180 GeV. Nevertheless, across this
0.8 cr[pb]
0.6
O.L.
0.2
ilE {GeV]
Figure 1. The excitation curve for top production near the threshold; Ref.[18] .
range the resonance remnants induce a steep rise of the cross section near
the threshold.
Since the rapid t decay restricts the interaction region to small distances,
the excitation curve can be predicted in perturbative QCD [17], based es-
sentially on the Coulombic interquark potential V(R) = -4/3 x Qs(R)/R.
The cross section is built up by the superposition of all nS(tl) states. The
form and the height of the excitation curve are very sensitive to the mass
of the top quark, Fig. l.
Detailed experimental simulations predict the following sensitivity to
the top mass [19] near mt rv 175 GeV:
omt ~ 200 MeV
for an integrated luminosity of J £ = 50 fb -1, remnant uncertainties due

to higher-order QCD corrections and experimental errors contributing at
approximate equal strength. At proton colliders a sensitivity of about 2 Ge V
has been claimed for the top mass, based on the reconstruction of top
quarks from jet and lepton final states. Thus, e+ e- colliders will improve
the measurement of the top quark mass by at least an order of magnitude.
72 P.M. ZERWAS
3. Electroweak Gauge Bosons

3.1. STANDARD W, Z BOSONS
The fundamental electroweak and strong forces appear to be of gauge the-

oretical origin. This is one of the outstanding results of theoretical and
experimental analyses in the past three decades. However, little direct evi-
dence has been accumulated so far for the non-abelian nature of the forces
in the electroweak W±, Z, "( sector. Since deviations from the gauge sym-
metries manifest themselves in experimental observables with coefficients
(Jh)2, high energies will allow stringent direct tests of the self-couplings of
the electroweak gauge bosons.
The gauge symmetries of the Standard Model determine the form and
the strength of the self-interactions of the electroweak bosons, triple cou-
plings WW,,(, WWZ and quartic couplings. Deviations from the gauge sym-
metry form of these vertices could be expected in more general scenarios
[20]. In models in which W, Z bosons are generated dynamically and inter-
act strongly with each other at high scales A, corrections could alter the
vertices to order {Mw / A)2 and induce new types of couplings.
While the experimental analyses of the self-couplings of the electroweak
bosons can be carried out at collider energies of 500 Ge V with high accuracy,
WW scattering [21] can only be studied at energies in the TeV range. This
is an important process which must be investigated very thoroughly if light
Higgs particles do not exist and W bosons become strongly interacting
particles at high energies.
The properties of the Z boson have been studied at LEP and SLC with
very high accuracy. By operating TESLA at low energies on the Z reso-
nance in the GigaZ mode and near the WW threshold, the measurement of
fundamental electroweak parameters, mixing angle and W mass can be im-
proved by yet an order of magnitude, allowing to test electroweak symmetry
breaking stringently at the quantum level.
The GigaZ Mode: By building a bypass for the transport of electron and
positron bunches, high luminosity can also be reached at low energies. On
the Z resonance, an ensemble of 109 events, 1 GigaZ, can be generated
in a year, expanding the LEP sample by two orders of magnitude. With
both electron and positron beams longitudinal polarized, the electroweak
mixing angle can be determined very accurately by measuring the left-right
asymmetry [22]
Similarly the measurement of the W mass can be improved to

oMw ~ 6MeV
by scanning the threshold region [23].
Based on these two measurements, many high-precision tests can be
performed in the electroweak sector. Extracting the Higgs mass from the
electroweak observables is particularly interesting. From the p--parameter
[26] this mass can be predicted to an accuracy of about 6%, improving
LHC based predictions by almost an order of magnitude [24]. Comparing
this prediction with the direct measurement of the Higgs mass, quantum
fluctuations can be tested stringently in a spontaneously broken gauge the-
ory.
The Triple Gauge Boson Couplings: The couplings W+W-,), and W+W- Z
are in general described each by seven parameters. Assuming C, P and T
invariance in the pure electroweak boson sector, the number of parameters
can be reduced to three,
r /
'-'j gj = .
zgj
lW*1.tV WAh
f.L v + .c. + . W*W
Zfijf.L v Ff.LV + . M2
Z
Aj W*Pf.L W /-LV F.VP
w
with g, = e and gz = e cot Ow for j = ,)" Z. The", = 1 + D.", and the A
parameters can be identified with the ,)" Z charges of the W bosons and
the related magnetic dipole moments and electric quadrupole moments,
J..l, = 2~w [2 + D."" + A,] and 'Y Z

~
Q, - ~?v [1 + D."" - A,] and 'Y ~ Z
The gauge symmetries of the SM demand '" = 1 and A = O. The mag-

netic dipole and the electric quadrupole moments can be measured directly
through the production of W'Y and W Z pairs at pp/pp colliders and WW
pairs at e+ e- and 'Y'Y colliders.
Detailed experimental analyses have been carried out for the reaction
e+e- ~ W+W- ~ (lve)(qq'). The bounds on D."" A which can be obtained
at e+e- colliders of 500 GeV [25] are significantly better than the bounds
expected from the LHC:
0"', = 4.8 X 10- 4 O"'z = 7.9 X 10- 4
A, = 7.2 X 10- 4 OAZ = 6.5 x 10- 4
ogf = 2.5 X 10- 3
Moreover, they improve at 1 TeV by nearly an order of magnitude. The
scales A which can be probed, extend far beyond the energy scales which
are accessible directly.
74 P.M. ZERWAS
Strongly Interacting W, Z Bosons If the scenario in which W, Z, Higgs

bosons are weakly interacting up to the GUT scale is not realized in Na-
ture, the alternative scenario is a strongly interacting W, Z sector. With-
out a light Higgs boson with a mass of less than about 1 Te V, the elec-
troweak bosons must become strongly interacting particles at energies of
about 1. 2 TeV to comply with the requirements of unitarity for the W L W L
scattering amplitudes. Based on Goldstone particles associated with the
spontaneous symmetry breaking, the new strong interactions may build up
the longitudinal degrees of freedom for the massive vector bosons in techni-
color type theories, for instance. In such scenarios, novel resonances could
be realized in the 0(1 TeV) energy range, generated in WLWL collisions.
In scenarios of strongly interacting vector bosons, W L W L scattering
must be studied at energies of order 1 TeV which requires the highest ener-
gies possible in e+e- colliders. (Quasi)elastic WW scattering can be inves-
tigated by using W bosons radiated off the electron and positron beams,
ee -7 vvWW, or by exploiting final state interactions in the e+e- annihi-
lation to W pairs, e+e- -7 W+W-. All possible (isospin, angular momen-
tum) combinations in WW scattering amplitudes aIJ can be realized in the
first process. The cross sections however are small as long as no resonances
are formed.
Building up the electroweak vector boson masses by the interactions of
the gauge fields with the Goldstone bosons associated with the spontaneous
symmetry breaking of the underlying strong-interaction theory, the longi-
tudinal degrees of freedom of the vector bosons can be identified at high
energies with the Goldstone bosons themselves as a result of the equiva-
lence theorem. In analogy to the 7r7r low-energy theorems, the first terms in
the energy expansion of the WW scattering amplitudes [27] are determined
independent of dynamical details:
6
aoo = + 967rv 2
1
all = + 967rv 2
These fundamental scattering amplitudes in the threshold region of the

strong WW interactions can be tested to an accuracy of about 15% at a
high-luminosity collider of 1 TeV [28J.
The attractive I = 0 and I = 1 channels may form Higgs and p-type
resonances at high energies. The formation of resonances would lead to
spectacular phenomena in WW collisions [29].
Similar phenomena would also be observed as rescattering effects in the
cross section O"(e+e- -7 W+W-) for W -pair production. (1, J) = (1,1)
resonance effects would be noticeable at Vs = 1 Te V up to resonance
masses of about 5 Te V in the angular distributions of the W decay final

states [30].
3.2. EXTENDED GAUGE THEORIES
The gauge symmetry of the Standard Model, SU(3) x SU(2) x U(I), is

widely believed not to be the ultima ratio. The SM does not unify the elec-
troweak and strong forces since the coupling constants of these interactions
are different and appear to be independent. However, one should expect
that in a more fundamental theory the three forces are described within
a single gauge group and , hence, with only one coupling constant at high
energy scales. This grand unified theory will be based on a gauge group
containing SU(3) x SU(2) x U(I) as a subgroup and it will be reduced to
this symmetry at low energies.
Two predictions of grand unified theories may have interesting conse-
quences in the energy range of a few hundred GeV [31]:
(i) The unified symmetry group must be spontaneously broken at the
unification scale AGUT 2: 1016 GeV in order to be compatible with the
experimental bounds on the proton lifetime. However, the breaking to the
SM group may occur in several steps and some subgroups may remain
unbroken down to a scale of order 1 Te V. In this case the surviving group
factors allow for new gauge bosons with masses not far above the scale
of electroweak symmetry breaking. Besides SU(5), two other unification
groups have received much attention: In SO(1O) three new gauge bosons
wj, ZR are predicted, while in E(6) a light neutral Z' boson may exist in
the Te V range.
The virtual effects of a new ZR , Z' boson associated with the most
general effective theories which arise from breaking
E(6) -+ SU(3) x SU(2) x U(l) x U(l)y, and SO(lO) -+ SU(2)L x SU(2)R x U(l),
have been investigated in Ref. [32].
Assuming the ZR, Z' bosons to be heavier than the available c.m. energy,
the propagator effects on various observables of the process
+ v -
e e- ~ ii: V = "(,Z and ZR,Z'
have been studied in detail. As shown in Table 1, the effects of new vector
bosons can be probed for masses up to 5 TeV at a 500 GeV collider. While
they may be produced directly up to about 5 TeV at the LHC, experiments
at the e+ e- collider will measure the couplings of the vector bosons to
fermions very precisely, thus identifying the physical nature of the new
bosons. Masses up to 10 Te V and 50 Te V can be probed in e+ e- colliders
operating at 800 Ge V and 5 Te V, respectively. These two windows extend
to much higher scales than the discovery limits anticipated at LHC.
76 P.M. ZERWAS
VB I SO(10) I E(6) I
500GeV 6TeV 5-7TeV
800GeV 10TeV 8-11 TeV
5TeV '" 50TeV ",50 TeV
TABLE 1. Sensitivity limits of ZR

masses in SO(10) and Z' masses in
E(6) at e+e- linear colliders in the
TeV range.
(ii) The grand unification groups incorporate extended fermion repre-

sentations in which a complete generation of 8M quarks and leptons can
be naturally embedded. These representations accommodate a variety of
additional new fermions. It is conceivable that the new fermions acquire
masses not much larger than the Fermi scale. This is necessary if the pre-
dicted new gauge bosons are relatively light. 80(10) is the simplest group in
which the 15 members of each 8M generation of fermions can be embedded
into a single multiplet. This representation has dimension 16 and contains
a right-handed neutrino. The group E(6) contains 8U(5) and 80(10) as
subgroups. In E(6), each quark-lepton generation belongs to a representa-
tion of dimension 27. To complete this representation, twelve new fields are
needed in addition to the 8M fermion fields. In each family the spectrum
includes two additional isodoublets of leptons, two isosinglet neutrinos and
an isosinglet quark with charge -1/3.
If the new particles have non-zero electromagnetic and weak charges,

and if their masses are smaller than the beam energy of the e+C collider,
they can be pair produced. In general, the production processes are built
up by a superposition of s-channel 'Y and Z exchanges, but additional con-
tributions could come from the extra neutral bosons if their masses are
not much larger than the c.m. energy [31]. The cross sections are large,
of the order of the point- like QED cross section. This leads to samples of
several thousands of events. Fermion mixing, if large enough, gives rise to
additional production mechanisms for the new fermions: single production
in association with the light partners. In this case, masses very close to the
total energy of the e+ e- collider can be reached.
4. Higgs Bosons
The Higgs mechanism is the cornerstone in the electroweak sector of the
Standard Model. The fundamental SM particles, leptons , quarks and weak
gauge bosons, acquire masses through the interaction with a scalar field. To
accommodate the well-established electromagnetic and weak phenomena,
the Higgs mechanism requires the existence of at least one weak iso-doublet
scalar field. After absorbing three Goldstone modes to build up the longi-
tudinal polarization states of the W±, Z bosons, one degree of freedom is
left-over, corresponding to a real scalar particle.
Three steps are necessary to establish the Higgs mechanism sui generis
experimentally as the mechanism for generating the masses of the funda-
mental SM particles:
(i) The Higgs boson must be discovered, the experimentum crucis.
(ii) It must be proven that the couplings of the Higgs particle with gauge
bosons and fermions increase with their masses;
(iii) The Higgs potential generating the non- zero Higgs field in the vacuum
and breaking the electroweak symmetry in the scalar sector must be
reconstructed by determining the Higgs self- couplings.
The only unknown parameter in the SM Higgs sector is the mass of the
Higgs particle. Constraints on the mass can, however, be derived from the
upper scale A of the energy range in which the model is assumed to be valid
before the particles become strongly interacting and new dynamical phe-
nomena emerge [33]. Increasing the energy scale, the quartic self-coupling
of the Higgs field grows logarithmically for large values indefinitely. If the
Higgs mass is small, the energy cut-off A is large at which the coupling
grows beyond any bound; conversely, if the Higgs mass is large, the cut- off
A is small. The condition MH < A sets an upper limit on the Higgs mass
in the Standard Model. Detailed analyses lead to an estimate of about
700 GeV for the upper limit on MH. If the Higgs mass is less than 180
to 200 GeV, the Standard Model can be extended up to the GUT scale
AGUT '" 10 16 GeV, while all particles remain weakly interacting. The hy-
pothesis that interactions between W, Z bosons and Higgs particles remain
weak up to the GUT scale, plays a key role in deriving the experimen-
tal value of the electroweak mixing parameter sin2 Ow from grand unified
theories. From this hypothesis and the additional requirement of vacuum
stability, upper and lower bounds on the Higgs mass can be derived. Based
on these arguments, the SM Higgs mass should be expected in the mass
window 130 < MH < 180 GeV for a top mass value of about 175 GeV.
A large variety of channels can be exploited to search for Higgs particles
in the Higgs-strahlung and fusion processes of e+e- colliders [34, 35, 36].
In the Higgs-strahlung process e+e- -+ ZH, missing-mass techniques can
78 P.M. ZERWAS
be used in events with leptonic Z decays or the Higgs particle may be

reconstructed in H -+ bb, WW directly. The WW fusion process e+ e-
-+ DeveH requires the reconstruction of the Higgs particle.
Once the Higgs boson is found at LEP, Tevatron or LHC, it will be very
important to explore its properties at the e+ e- linear collider to estab-
lish the Higgs mechanism sui generis experimentally. This is possible with
high precision in the clean environment of e+ e- colliders in which at high
luminosity a large ensemble of order 105 Higgs bosons can be generated
nearly background- free. The zero- spin of the Higgs particle is reflected in
the angular distribution of the Higgs-strahlung process which asymptoti-
cally must approach the sin2 () law. The strength of the couplings to Z and
W bosons is reflected in the magnitude of the e+ e- production cross sec-
tions. The strength of the couplings to fermions can be measured through
the decay branching ratios and Higgs bremsstrahlung off top quarks. Dou-
ble Higgs- strahlung can be exploited to measure the trilinear Higgs self-
coupling.
From the preceding discussion we conclude that an e+ e- linear collider
with energies in the range of 300 to 500 GeV and high luminosity is the ideal
instrument to investigate the Higgs mechanism in the intermediate mass
range which, a priori, may be considered the most important part in the en-
tire range of possible Higgs mass values. A similarly comprehensive analysis
can be performed in supersymmetric extensions of the Standard Model. In
particular, the Minimal Supersymmetric extension of the Standard Model
MSSM incorporates a spectrum of five Higgs particles (representative for a
wide class of models), which can be explored in all facets at e+e- colliders.
At tree level, the mass of the lightest Higgs boson hO is smaller than the Z
mass; the bound is shifted to'" 150 GeV by radiative corrections. This par-
ticle is guaranteed to be detected at e+ e- colliders - even in non-minimal
versions of the supersymmetric theories. The masses of the heavy neutral
and charged Higgs bosons can be expected in the range of the electroweak
symmetry breaking scale so that a large part of the mass parameter space
can be covered for the individual Higgs particles at e+ e- colliders operat-
ing at an energy of 1 TeV. The detailed comparison of branching ratios and
production cross sections with the predictions of the MSSM will eventually
shed light on the physical basis of the extended Higgs sector.
4.1. THE HIGGS PARTICLE IN THE STANDARD MODEL
The profile of the SM Higgs particle is completely determined if the Higgs

mass is fixed. For Higgs particles in the intermediate mass range Mz ::;
MH ::; 2Mz the main decay modes are decays into bb pairs and WW, ZZ
pairs with one of the two gauge bosons being virtual below the threshold
[37J. Above the WW threshold, the Higgs particles decay almost exclusively
into these channels, except in the mass range near the tI decay threshold.
Below 140 GeV, the decays H ~ 7+7-, cc and gg are also important besides
the dominating bb channel. Up to masses of 140 Ge V, the Higgs particle
is very narrow, f(H) ~ 10 MeV. After opening the [virtualJ gauge boson
channels, the state becomes rapidly wider, reaching rv 1 Ge V at the Z Z
threshold. The width cannot be measured directly in the intermediate mass
range. Only above MH ~ 200 GeV it becomes wide enough to be resolved
experimentally.
The main production mechanisms for Higgs particles in e+ e- collisions
are Higgs-strahlung off the Z boson line [38] and the WW fusion process
[39],
e+e- ~ Z+H
e+e- ~ lie Ve +H
With rising energy the Higgs-strahlung cross section scales rv a~/ s while
the fusion cross sections increase logarithmically rv a;MH;.2log s / M'k, be-
coming dominant above 500 Ge V:
G}Mt [1 + (1 - 4 sin2 OW)2J

967rs
G}Mtv I S
4V27r3 og M'k
As a general rule, the cross sections and rates [about 105 events] are suffi-
ciently large to detect Higgs particles with masses up to 70% of the total
e+ e- c.m. energy.
The recoiling Z boson in the two- body reaction e+e- ~ ZH is mono-
energetic and the mass can be derived from the energy of the Z boson,
M'k = s - 2y'sEz + M~. Initial state bremsstrahlung and beamstrahlung
smear out the peak slightly, as shown in Fig. 2. A similarly clear peak can
be observed in the fusion process e+e- ~ veveH by collecting the decay
products of the Higgs boson.
Mass and Width: The mass of the Higgs boson can be measured very
accurately by analyzing the recoil Z spectrum in Higgs-strahlung events.
Experimental simulations [41] have demonstrated that the error in the mass
measurement can be reduced to
80 P.M. ZERWAS
+ Data
~ ZH-HeX
m H = 120 GeV
o+--r--r--""~~=
100 120 140 160
Recoil Mass [Ge V]
Figure 2. Production of Higgs particles III the Higgs-strahlung channel

e+e- --+ ZH --+ e+e- Xj Ref. [40J.
in high-luminosity runs.
The width of the SM Higgs boson can be determined in an almost com-
pletely model-independent way in the difficult intermediate mass range in
which the Breit-Wigner form cannot be reconstructed at an e+e- collider.
Measuring the branching ratio B~ in the decay and the partial width ri
in the production process, the total width r H can be derived from
rH = rdBRi
The two channels i = WW [42,43] and i = 'Y'Y [44] are useful for this analy-
sis [45]. The partial width rww can be extracted from the size of the WW
fusion cross section [46] while r n can be measured in the Compton collider
mode [47]. The accuracies of a few percent match the expected accuracy in
scanning the Breit-Wigner excitation in a muon-collider.
Spin and Parity: The angular distribution of the Z I H bosons in the Higgs-
strahlung process is sensitive to the external quantum numbers of the Higgs
particle [38]. Since the amplitude is given by A(O+) ,...., Cl . C2, the Z boson
is produced in a state of longitudinal polarization at high energies. As a
result, the angular distribution dO" Id cos 0 ,...., ). sin2 0 + 8Mi Is approaches
the spin- zero law sin2 0 asymptotically. This may be contrasted with the
distribution,...., 1 + cos 2 0 for negative parity states, which follows from the
transverse polarization amplitude A(O-) '" C1 X C2' kz . It is also characteris-

tically different from the background process e+ e- -+ Z Z which is strongly
peaked in the forward/backward direction.
Higgs Couplings: Since the fundamental SM particles acquire masses through
the interaction with the Higgs field, the scale of the Higgs couplings to
fermions and gauge bosons is set by the masses of the particles:
and
It will be a very important task to measure the Higgs couplings to the fun-
damental particles [42, 43] since they are uniquely predicted by the very na-
ture of the Higgs mechanism. The Higgs couplings to massive gauge bosons
can be determined from the measurement of the production cross sections
with an accuracy of ±1%, the HZZ coupling in the Higgs-strahlung and
the HWW coupling in the fusion process. For Higgs couplings to fermions,
either loop effects in H ~ gg, 'Y'Y [mediated by top quarks] can be exploited,
or the direct measurement of branching ratios H -+ bb, ee, T+T-, gg in the
lower part of the intermediate mass range. This is exemplified in Fig. 3.
For MH = 120 GeV the following accuracy fJBR/ BR can be achieved [43]
in the determination of the Higgs decay branching ratios:
bb: 2.4% WW*: 5.4%

ee: 8.3% gg : 5.5%
TT: 6.0%
By measuring the ratio of the TT to the bb branching ratios
BR(H -+ TT) m;
BR(H -+ bb) - 3m~(MH)
the linear dependence of the Yukawa couplings on the fermion masses can
be tested very nicely. A direct way to determine the Yukawa coupling of the
intermediate mass Higgs boson to the top quark in the range mH ~ 120 GeV
is provided by the bremsstrahlung process e+ e- -+ tIH in high energy
e+e- colliders [48]. The absolute values of the Yukawa couplings can be
reconstructed by combining decay branching ratios with the production
cross sections.
Higgs Self- couplings: To generate a non- zero value of the Higgs field in
the vacuum, the minimum of the Higgs potential must be shifted away from
the origin. Rewriting the potential
82 P.M. ZERWAS
.2
iii bb
a:
Ol
c
J::
"til
C
cD
<I)
Ol
Ol
I
.,
10
TT
99
CC
Figure 3. Branching ratios of SM Higgs decays into fermion and WW· pairs; Ref.[43J.
in terms of the physical Higgs field H, the potential can be reconstructed by

measuring the trilinear and quartic couplings. At a high- luminosity e+ e-
collider, the trilinear coupling can be tested in the double Higgs- strahlung
process
The splitting of a virtual Higgs boson into two real Higgs bosons is de-
termined by the trilinear Higgs coupling: e+e- -t Z + H* [-t HH]. Even
though the cross section is less than 1 fb [49], the coupling can be expected
nevertheless to be measured with an accuracy better than 20% [50]. Thus
an essential element of the mechanism responsible for the spontaneous sym-
metry breaking in the scalar sector can be established experimentally at the
high-luminosity collider.
PHYSICS WITH AN e+ e- LINEAR COLLIDER 83
4.2. SUSY HIGGS PARTICLES
The Higgs spectrum in the minimal supersymmetric extension of the Stan-

°
dard Model [51] consists of five states: hO, HO, A and H±. Besides the
masses, two mixing angles define the properties of the scalar particles and
their interactions with gauge bosons, fermions and the self-interactions: the
ratio of the two vacuum expectation values tgj3 = V2/Vl and a mixing angle
0: in the neutral CP-even sector. Supersymmetry leads to several relations
among the masses and mixing angles, and, in fact, only two parameters are
independent.
Neutral Higgs Bosons: The lightest neutral Higgs boson will decay mainly
into fermion pairs since its mass is smaller than 150 GeV. This is also the
f'oJ
dominant decay mode of the pseudoscalar boson A 0 . For values of tgj3 larger
than unity and for masses less than f'oJ140 GeV, the main decay modes of
the neutral Higgs bosons are decays into bb and 7+7- pairs; the branching
ratios are of order 90% and 8%, respectively. The decays into cc pairs
f'oJ
and gluons [proceeding through t and b quark loops] are suppressed, for
large tgj3 strongly. For large masses, the top decay channels HO, AO -t tl
open up; yet this mode remains suppressed for large tgj3. For large tgj3,
the neutral Higgs bosons decay almost universally into bb and 7+7- pairs.
If the mass is high enough, the heavy CP- even Higgs boson HO can in
principle decay into weak gauge bosons, HO -t WW, ZZ. Since the partial
widths are proportional to cos 2 (j3 - 0:), they are strongly suppressed and
the gold-plated Z Z signal of the heavy SM Higgs boson is lost in the
supersymmetric extension. The heavy neutral Higgs boson HO can also
decay into two lighter Higgs bosons. These modes, however, are restricted
to small domains in the parameter space. Other possible channels are decays
into supersymmetric particles. While sfermions are likely too heavy to affect
Higgs decays in the mass range considered here, Higgs boson decays into
charginos and neutralinos could eventually be important since the masses
of some of these particles are expected to be of order M z. [These new
channels are summarized in Ref.[52].] The charged Higgs particles decay
into fermions but also, if allowed kinematically, into the lightest neutral
Higgs boson and a W boson. Below the tb and W h thresholds, the charged
Higgs particles will decay mostly into 7Vr and cS pairs, the former being
dominant for tgj3 > 1. For large M H± values, the top-bottom decay mode
H+ -t tb becomes dominant.
Adding up the various decay modes, the width of all five Higgs bosons
remains very narrow, being of order 10 GeV even for large masses.
84 P.M. ZERWAS
The search for the neutral SUSY Higgs bosons at e+ e- colliders will be
a straightforward extension of the search performed at LEP2 . This collider
is expected to cover the mass range up to 90 to 100 GeV for neutral Higgs
"-J
bosons, depending on tgj3 . Higher energies, ...;s 250 GeV, are required
"-J
to sweep the entire parameter space of the MSSM . The main production
mechanisms of neutral Higgs bosons at e+e- colliders [52,51) are the Higgs-
strahlung process and associated pair production,
(a) Higgs - strahlung e+e- -7 Z+hIH

(b) pair production e+ e- -7 A + hi H
as well as the related fusion processes. The CP-odd Higgs boson AO cannot
be produced in fusion processes since it does not couple to gauge bosons in
leading order.
The cross sections for the four Higgs-strahlung and pair production
processes can be expressed as
a(e+e- -7 Zh) = sin2 (j3 - a) aSM
a(e+e- -7 ZH) cos 2(j3 - a) aSM
and
2 -
a(e+ e- -7 Ah) cos (13 - a) A aSM
2 -
a(e+e- -7 AH) sin (13 - a) A aSM
where aSM is the SM cross section for Higgs-strahlung and the factor X
accounts for the suppression of the P-wave cross sections near the thresh-
old. The cross sections for the Higgs-strahlung and for the pair production
as well as the cross sections for the production of the light and the heavy
neutral Higgs bosons hO and HO are mutually complementary to each other,
coming either with a coefficient sin2 (j3 - a) or cos 2(j3 - a). As a result, since
aSM is large, at least the lightest CP-even Higgs boson must be detected.
The cross section for hZ in the Higgs-strahlung process is large for
values of Mh near the upper bound. The heavy CP-even and CP-odd
Higgs bosons H and A, on the other hand, are produced pairwise in this
limit:
The decoupling limit corresponds to heavy Higgs masses above 250 to

300 GeV. The discovery limit is therefore set by the beam energy indepen-
dently of the mixing parameter tgj3, in contrast to LHC where the heavy
Higgs bosons cannot be detected individually in parts of the parameter
space.
Charged Higgs Bosons: The charged Higgs bosons, if lighter than the top
quark, can be produced in top decays, t -+ b + H+, with a branching ratio
varying between 2% and 20% in the kinematically allowed region. Since for
tg,8 larger than unity, the charged Higgs bosons will decay mainly into TVT ,
this results in a surplus of T final states over e,11 final states in t decays,
an apparent breaking of lepton universality. For large Higgs masses the
dominant decay mode is the top decay H+ -+ tb. In this case the charged
Higgs particles must be pair produced in e+ e- colliders:
e+e- -+ H+ H-
The cross section depends only on the charged Higgs mass. For small Higgs
masses the cross section is of order 100 fb at ..jS = 500 GeV, but it drops
very quickly due to the P-wave suppression'" ,83 near the threshold.
The preceding discussion of the MSSM Higgs sector at e+e- linear col-
liders can be summarized in the following two points:
(i) The lightest CP- even Higgs particle hO can be detected in the entire
range of the MSSM parameter space, either through the Higgs-strahlung
process e+e- -+ hZ or through pair production e+e- -+ hA [41]. In fact,
this conclusion holds true even at a c.m. energy of 250 GeV, independent
of the top and squark mass values, and also if invisible neutralino decays
are allowed for.
(ii) The area in the parameter space where all SUSY Higgs bosons individu-
ally can be discovered at e+e- colliders is characterized by MA .:s !..jS, inde-
pendently of tg,8. The HO, AO Higgs bosons can be produced either through
Higgs- strahlung or through Ah, AH associated production; charged Higgs
bosons will be produced in H+ H- pairs up to the kinematical limit.
5. Supersymmetry
Even though there is no direct experimental evidence so far for the real-
ization of supersymmetry in Nature, this concept has so many attractive
features that it can be considered as a prime target of present and future
experimental particle research [53]. Arguments in favor of supersymmetry
are deeply rooted in particle physics. Supersymmetry may play an impor-
tant role in a quantum theory of gravity. In relating particles of different
spins to each other, i.e. fermions and bosons, low-energy supersymmetry
stabilizes the masses of fundamental Higgs scalars in the background of
very high energy scales associated with grand unification. Besides solving
this hierarchy problem, supersymmetry may even be closely related to the
physical origin of the Higgs phenomenon itself: In a supergravity inspired
GUT realization with universal scalar masses at the GUT scale, the evolu-
tion of one of the scalar masses squared can become negative and can thus
86 P.M. ZERWAS
give rise to spontaneous symmetry breaking if the top mass has a value
between about 100 to 200 GeV while all other squared masses of squarks
and sleptons remain positive so that U(1)EM and SU(3)c remain unbroken.
The minimal supersymmetric extension of the Standard Model is based

on the SM group SU(3) x SU(2) x U(1). The gauginos are the supersym-
metric spin-~ partners of the gauge bosons. The quark and lepton matter
particles are associated with scalar supersymmetric particles, squarks and
sleptons. To preserve supersymmetry, two Higgs doublets are needed to en-
dow down as well as up-type fermions with masses; the supersymmetric
partners of the Higgs bosons are spin-~ higgsinos. [Charged/neutral hig-
gsinos mix in general with the non-colored gauginos, forming charginos
and neutralinos.] Supersymmetric partners carry a multiplicative quantum
number R = -1 (R = +1 for ordinary particles) which is conserved in this
model. Supersymmetric particles are therefore generated in pairs and the
lightest supersymmetric particle LSP is stable.
Strong support for supersymmetry and the MSSM particle spectrum in
the mass range of several hundred Ge V follows from the high-precision mea-
surement of the electroweak mixing angle sin2 Ow [58]. The value predicted
by the MSSM and the value determined by the LEP and other experiments,
MSSM: sin2 Ow = 0.2336 ± 0.0017

EXP sin2 Ow = 0.2316 ± 0.0002
match surprisingly well, the difference being less than about 2 per-mille.
A central problem of supersymmetric theories is the breaking mecha-

nism. Several scenarios have been proposed and experimental consequences
have been elaborated in a few cases: gravity mediated supersymmetry
breaking mSUGRA [54]; gauge mediated supersymmetry breaking GMSB
[55]; anomaly mediated supersymetry breaking AMSB [56]; Scherk-Schwarz
supersymmetry breaking SSSB [57]. Mass spectra of the supersymmetric
particles are quite different in these scenarios so that high-precision mea-
surements of the particle properties will shed light experimentally on this
theoretical problem. Moreover, extrapolations can be performed in a stable
way which will allow to reconstruct the basic supersymmetric theory even-
tually at a scale close to the Planck scale. High-precision experiments at
a high-luminosity e+ e- linear collider are expected to advance the under-
standing of supersymmetry in an essential way.
5.1. SUPERSYMMETRIC PARTICLES
Charginos and Neutralinos The two charginos xi

and the four neutrali-
nos X?, mixtures of the [non-colored] gauginos and higgsinos, are expected
to be the lightest supersymmetric particles. In the MSSM with conserved
R-parity, the neutralino X~ with the smallest mass, assumed to be the
lightest supersymmetric particle, is stable. The heavier neutralinos and the
charginos decay into (possibly virtual) gauge and Higgs bosons plus the
LS P, X? -+ X~ + Z and X~ + W, or if they are heavy enough, into neu-
tralino/chargino cascades, and leptons plus sleptons [59].
Neutralinos and charginos are easy to detect at e+ e- colliders. They
are produced in pairs
e+e- -+ xi + x; [i,j = 1,2]

e+e- -+ x? + xJ [i, j = 1, .. ,4]
through s-channel "Z exchange and t-channel select ron or sneutrino ex-
change. Since the cross sections are as large as 0(100 fb), enough events
will be produced to discover these particles for masses nearly up to the
kinematical limit.
The properties of the charginos and neutralinos can be studied in great
detail at e+ e- colliders. From the fast onset '" j3 of the spin - ~ excita-
tion curve near the threshold, the masses can be measured very accurately
within less than a hundred MeV. The decay spectra of heavier supersym-
metric particles allow us to measure the mass of the LSP within ± 2
GeV, Fig. 4. Using polarized e± beams, the decomposition of the states,
xi = aW+ + j3fI+ into wino and higgsino components can be determined
[61] since left-handed electrons couple to sneutrinos in the t-channel but
right-handed electrons do not, so that the energy and angular dependence
of the cross sections is different for the two polarization states [62]. In a
similar way the properties of neutralinos can be explored [63].
Sleptons and Squarks: The superpartners of the right-handed leptons de-

cay into the associated SM partners and neutralinos/charginos. In major
parts of the SUSY parameter space the dominant decay mode is jJ,R -+
I-l + X~ [59]. For the superpartners of the left-handed sleptons, the decay
pattern is slightly more complicated since, besides the X~ channels, decays
into leptons and charginos are also possible. In e+ e- and e- e- collisions,
sleptons are produced in pairs:
-+ -+-- -+-- -+-- -+--

I-l L I-l L , I-l RI-l R ' TL TL , TR TR
-+ ihiiL
88 P.M. ZERWAS
200 X~ X;
• l. = 10 fb - I/polli.l
150
800
100
400
50
20 40 60 80 100
lepLon e n e rgy 1':, [CcV]
o ~~--~~~~~~~~
260 264 268

Eom. [CeV)
t
Figure 4. The excitation curve for chargino production e +e - ~ X x"1 near the threshold
and energy distribution of the final state /-t in the decay ilR ~ /-t + X~ in flight; Ref.[60].
--+ -+-- -+-- -+--

eLe L ' eRe R , eLe R , eRe L
-+--
e e --+ eLe L ' eRER , eLeR
For charged sleptons, the production proceeds via 'Y, Z exchange in the s-
channel, in the case of selectrons, also by additional t-channel neutralino
exchange. For sneutrinos, the process is mediated by s-channel Z- exchange
and, in the case of electron-sneutrinos, also by the t-channel exchange of
charginos.
The cross sections for the pair production of sleptons are of the order of
10- 1 to 10- 2 pb so that their discovery is very easy up to the kinematical
limit. From the P-wave onset {33 of the annihilation cross section the
tv
masses can in general be determined [60] at a level of 200 to 300 MeV; the
sharper onset of selection production in e- e- scattering [64] will reduce
this number further. Enough events will be produced to study their de-
tailed properties. The polarization of the e± beams will help to identify the
couplings of these particles.
If one of the stop states is light enough due to the strong LR Yukawa
mixing, these particles may be pair produced even at a 500 GeV collider:
i,j = 1,2
By measuring the LR asymmetry ofthe production cross sections, the h/tR

mixing angle can be determined to high accuracy 0 cos(h ~ 0.01 [65].
5.2. TESTING SUSY -GUT
The high precision with which masses, couplings and mixing parameters
will be determined at e+ C colliders, can be exploited to test the mecha-
nism for supersymmetry breaking and the structure of underlying theories.
In minimal supergravity mSUGRA the breaking of supersymmetry is me-
diated by gravity from a hidden sector to the eigenworld, generating soft
SUSY breaking parameters at the grand unification scale 1. The parameters
are generally assumed to be universal at that scale in the gaugino and the
scalar sectors. In gauge mediated supersymmetry breaking, gauge interac-
tions connect the mechanism to the eigenworld at a scale possibly between
10 and 105 TeV. Mass spectra in mSUGRA and GMSB are characteristi-
cally different , the splitting between sleptons and squarks being larger in
GMSB. Moreover, since the gravitino in GMSB is very light, X~ or 71 may
be long lived giving rise to displaced photons or stable heavy tracks in the
decays X~ -* ,(; or 71 -* T(; with the gravitino (; escaping undetected [66J.
If minimal supergravity is the underlying theory, the observable prop-

erties of the superparticles can be expressed by a small set of parameters
defined at the GUT scale. Basically four parameters specify the properties
of the particles in the supersymmetric sector. The scalar mass parameter
mo, the SU(2) gaugino mass M 1 / 2 , the trilinear coupling Ao and the sign
of the coupling f-L of the Higgs doublets in the superpotential, and tan,B,
the ratio of the vacuum expectation values VdVl. The Higgs sector requires
another parameter which is in general identified with the mass of the pseu-
doscalar Higgs boson MA. Evolving the scalar masses from the GUT scale
down to low energies, it turns out that non- colored particles are in general
significantly lighter than colored particles. The lightest of the non-colored
gauginos/higgsinos and sleptons could have masses in the range of 100 to
200 GeV. Since only a few parameters determine the low energy theory of
the evolution from the GUT scale down to the electroweak scale, many
1 If gravitational interactions would become strong not at very high scale but near the
electroweak scale [67], collider experiments could probe the additional spatial dimensions
through which gravitational fields could propagate [68]. Contact interactions and missing
energy events could signal Planck scales in 4 + n dimensions up to about 10 TeV. Thus,
the basic space-time structure can be explored in these experiments.
90 P.M. ZERWAS
relations can be found among the masses of the superparticles which can
stringently be tested at e+ e- colliders. Two examples should illustrate the
physics potential of the e+ e- facilities in this context.
(i) The gaugino masses at the scale of SU2 x U1 symmetry breaking are
related to the common SU2 gaugino mass M 1/ 2 at the GUT scale by the
running gauge couplings:
Mi = -O:i
- M1/ 2 i = SU3, SU2, U1
O:GUT
with O:GUT being the gauge coupling at the unification scale. The mass
relation in the non-color sector
Ml 5 2 1
- = -tan Ow ~
M2 3 2
can be tested very well by measuring the masses and production cross
sections of charginos/neutralinos and sleptons.
(ii) In a similar way the slepton masses can be expressed in terms of a
common scalar mass parameter mo at the GUT scale, contributions", M 1/ 2
due to the evolution from the GUT scale down to low energies, and the D
terms related to the electroweak symmetry breaking. These expressions give
rise to relations among the slepton masses:
-(1 - 2 sin2 Ow) cos 2t3m~

"'L Ml/2 - HI - 4 sin2 Ow) cos 2t3m~
with "'L = 0.37. The second relation follows from the hypothesis that the
scalar masses are universal at the GUT scale, in particular m 2 (5*) = m 2 (1O)
within SU5. This assumption can be tested by relating the mass difference
between eL and eR to the SU2 gaugino mass.
The typical results of an overall fit are illustrated in Table 2 [69].
In a much more difficult procedure the parameters may not only be
fitted by assuming a universal set at the GUT scale from the start, but
the set itself may be reconstructed by evolving the mass parameters from
the electroweak scale to the unification scale [69]. This is the only way to
reconstruct operationally the fundamental supersymmetric theory near the
Planck scale from experimental observations at the electroweak scale.
Precision tests of supersymmetric particles in e+ e- collider experiments

can thus open a window to energy scales close to the Planck scale where
gravity, the fourth of the fundamental forces, becomes an integral part of
physics.
I Parameter I Theor. Value I Meas. Error I

mo 100GeV 0.09GeV
M 1/ 2 200GeV O.lOGeV
Ao OGeV 6.30GeV
tan{3 3 0.02
TABLE 2. Measurement of universal

mSUGRA parameters.
References
1. S.L. Glashow, Nucl. Phys. 20 (1961) 579; S. Weinberg, Phys. Rev. Lett. 19 (1967)
1264; A. Salam, in Elementary Particle Theory, ed. N. Svartholm (Almqvist and
Wiksells, Stockholm 1968).
2. H. Fritzsch and M. Gell-Mann, Proc. XVI Int. Conf. on High Energy Physics, eds .
J.D. Jackson and A. Roberts, eds, (Fermilab 1972).
3. P.W. Higgs, Phys. Rev. Lett. 12 (1964) 132; Phys. Rev. 145 (1966) 1156; F. Englert
and R. Brout, Phys. Rev . Lett. 13 (1964) 321; G.S. Guralnik, C.R. Hagen and
T.W. Kibble, Phys. Rev. Lett. 13 (1964) 585.
4. J.C. Pati and A. Salam, Phys. Rev. DI0 (1974) 275; H. Georgi and S. Glashow,
Phys. Rev. Lett. 32 (1974) 438; H. Fritzsch and P. Minkowski, An . Phys. 93 (1975)
193.
5. J. Wess and B. Zumino, Nucl. Phys. B70 (1974) 39.
6. E . Accomando et a!. , Phys. Rep. 299 (1998) 1, [hep-ph/9705442] .
7. H. Murayama and M.E. Peskin, Ann. Rev. Nuc!. Part . Sci. 46 (1996) 553; P.M. Zer-
was, Surv. High Energy Phys. 12 (1998) 209.
8. F . Richard, Proceedings, LC Conference , Sitges 1999, [hep-ph/ 9909317].
9. ATLAS Collaboration, Technical Design Report 1999; CMS Collaboration, Physics
with CMS [www page].
10. Prospective Study of Muon Storage Rings at CERN, B. Autin, A. Blondel and
J. Ellis, eds, CERN 99- 02; C.M. Ankenbrandt et a!., Phys. Rev. ST Acce!. Beams
2 (1999) 081001.
11. F. Abe et al. [CDF Coli.], Phys. Rev. D50 (1994) 2966 and Phys. Rev. Lett. 74
(1995) 2626; S. Abachi et al. [DO Coli.], Phys. Rev. Lett. 74 (1995) 2632.
12. 1.1. Bigi, Y.L. Dokshitzer, V. Khoze, J . Kuhn and P.M. Zerwas, Phys. Lett. 181B
(1986) 157.
13. O. Podobrin, Proceedings, e+e- Collisions at 500 GeV: The Physics Potential,
Munich-Annecy-Hamburg 1992/93, DESY 93- 123C.
14. J . Jersak, E. Laermann and P.M . Zerwas, Phys. Rev. D25 (1982) 363.
15. M.E. Peskin and C. Schmidt, in Proceedings, Physics and Experiments with e+e-
Linear Colliders, Saariselkii 1991, R. Orava, P. Eerola and M. Nordberg eds., World
92 P.M. ZERWAS
Scientific 1992.
16. W . Bernreuther et a!., in Proceedings, e+e- Collisions at 500 GeV: The Physics
Potential, Munich-Annecy-Hamburg 1991/93, DESY 92-123A+B, 93- 123C.
17. V.S. Fadin and V.A. Khoze, JETP Lett. 46 (1987) 525 and Sov. J. Nucl. Phys. 48
(1988) 309; A.H. Hoang and T . Teubner, DESY 99-047, [hep-ph/99044681
18. P. Igo-Kemenes, M. Martinez, R. Miquel and S. Orteu, in Proceedings, e e- Col-
lisions at 500 Ge V: Th e Physics Potential, Munich-Annecy-Hamburg 1991/93,
DESY 92-123A+B, 93-123C.
19. M. Martinez, Proceedings, LC Conference, Sitges 1999.
20. K. Hagiwara, R.D. Peccei and D. Zeppenfeld, Nucl. Phys. B282 (1987) 253.
21. B.W. Lee, C. Quigg and H.B. Thacker, Phys. Rev. Lett. 38 (1977) 883, and Phys.
Rev. D16 (1977) 1519.
22. R. Hawkings and K. Monig, EPJdir. C in prep., hep-ex/9910022 .
23. G. Wilson, Report at the DESY /ECFA LC Workshop, Strasbourg-Obernai 1999.
24. S. Heinemeyer, T . Mannel and G. Weiglein, DESY99-117, [hep-ph/ 9909538]; J . Er-
ler, S. Heinemeyer, W. Hollik, G. Weiglein and P.M. Zerwas, DESY 99-177.
25 . K. Monig, Report at the DESY /ECFA LC Workshop, Strasbourg-Obernai 1999.
26. M. Veltman, Acta Phys. Polon. B8 (1977) 475.
27. T . Appelquist and C. Bernard, Phys. Rev. D22, 200 (1980); A. Longhitano, Phys.
Rev. D22 , 1166 (1980); Nuc!. Phys. B188, 118 (1981) ; T . Appelquist and G.-H . Wu,
Phys. Rev. D48, 3235 (1993); M. Veltman, Report CERN 97- 05.
28. E. Boos, H.-J. He, W. Kilian, A. Pukhov, C.-P. Yuan and P.M. Zerwas, Phys. Rev.
D57, 1553 (1998) and hep- ph/9908409.
29. V. Barger, K. Cheung, T. Han and R.J . Phillips, Phys. Rev. D52 (1995) 3815.
30. T . Barklow, in Proceedings, Physics and Experiments with e+e- Linear Colliders,
Saariselka. 1991, R . Orava, P. Eerola and M. Nordberg eds. , World Scientific 1992;
D. Dominici, Riv. NOllV. Cim. 20 (1997) 1.
31. A. Djouadi, Proceedings, Workshop on Physics and Experiments with e+e- Linear
Colliders, Iwate 1995.
32. A. Leike, Phys. Rep. 317 (1999) 143.
33. N. Cabibbo, L. Maiani, G. Parisi and R. Petronzio, Nucl. Phys. B158 (1979) 295;
M. Sher, Phys. Rep. 179 (1989) 273; M. Lindner, Z. Phys. C31 (1986) 295; J. Casas,
J. Espinosa and M. Quiros, Phys. Lett. B342 (1995) 171; K. Riesselmann, DESY 97-
222, [hep-ph/9711456].
34. J . Ellis, M.K. Gaillard and D.V. Nanopoulos, Nucl. Phys. B106 (1976) 292; B.L. Ioffe
and V.A. Khoze, Sov. J. Part. Nucl. 9 (1978) 50; B.W. Lee, C. Quigg and
H.B. Thacker, Phys. Rev. D16 (1977) 1519.
35. D.R.T. Jones and S.T. Petcov, Phys. Lett. B84 (1979) 440.
36. P.M. Zerwas, Acta Phys. Po!. B30 (1999) 1871; M. Spira and P.M. Zerwas, Pro-
ceedings, 36. Int. Universitatswochen, Schladming 1997, hep-ph/9803257; C. Quigg,
FERMILAB-CONF-99/033-T .
37. A. Djouadi, M. Spira and P.M. Zerwas, Z. Phys. C70 (1996) 427.
38. V. Barger, K. Cheung, A. Djouadi, B.A. Kniehl and P.M. Zerwas, Phys. Rev. D49
(1994) 79.
39. W . Kilian, M. Kramer and P.M. Zerwas, Phys. Lett. B373 (1996) 135.
40. P. Garcia-Abia and W. Lohmann, Proceedings, LC Conference, Sitges 1999, [hep-
ex/9908065].
41. P. Janot, in Proceedings, Physics and Experiments with e+e- Linear Colliders,
Waikoloa/Hawaii 1993, F. Harris, S. Olsen, S. Pakvasa and X. Tata eds., World
Scientific 1993.
42. G. Borisov and F. Richard, Proceedings, LC Conference, Sitges 1999, hep-
ph/9905413.
43. M. Battaglia, ibid., hep- ph/9910271.
44. D. Reid, Report at the DESY /ECFA Workshop, Oxford 1999.
45. W. Kilian, M. Miihlleitner and P.M. Zerwas, DESY 99-171.
46. S. Yamashia, Proceedings, LC Conference, Sitges 1999.

47. G. Jikia and S. Soldner-Rembold, Proceedings, LC Conference, Sitges 1999.
48. A. Djouadi, J. Kalinowski and P.M. Zerwas, Z. Phys. C54 (1992) 255; S. Dawson
and L. Reina, Phys. Rev. D59 (1999) 054012; S. Dittmaier, M. Kramer, Y. Liao,
M. Spira and P.M. Zerwas, Phys. Lett. B 441 (1998) 383.
49. A. Djouadi, W. Kilian, M. Miihlleitner and P.M. Zerwas, Eur. Phys. J. ClO (1999)
27.
50. P. Lutz and P. Gay, Reports at the DESY /ECFA Workshops, Oxford/Strasbourg-
Obernai 1999.
51. J.F. Gunion and H.E. Haber, Nucl. Phys. B272 (1986) 1 and B278 (1986) 449.
52. A. Djouadi, J. Kalinowski and P.M. Zerwas, Z. Phys. C57 (1993) 569; A. Djouadi,
J. Kalinowski, P. Ohmann and P.M. Zerwas, Z. Phys. C74 (1997) 93.
53. F. Zwirner, Proceedings, 36. Int. Universitatswochen, Schladming 1997.
54. A.H. Chamseddine, R. Arnowitt and P. Nath, Phys. Rev. Lett. 49 (1982) 970;
R. Barbieri, S. Ferrara and C.A . Savoy, Phys. Let. B1l9 (1982) 343.
55. M. Dine, A. Nelson and Y. Shirman, Phys. Rev. D51 (1995) 1362; M. Dine, A. Nel-
son, Y. Nir and Y. Shirman, Phys. Rev. D53 (1996) 2658; G.F. Giudice and R. Rat-
tazzi, CERN Report, hep-ph/9801271.
56. G.F. Giudice, M.A . Luty, H. Murayama and R. Rattazzi, JHEP 9812 (1998) 027.
57. J. Scherk and J.H. Schwarz, Nucl. Phys. B153 (1979) 61.
58. J. Ellis, S. Kelley and D. V. Nanopoulos, Phys. Lett. 260B (1991) 131 ; U. Amaldi,
W. de Boer and H. Fiirstenau, Phys. Lett . 260B (1991) 447; P. Langacker and M.
Luo, Phys. Rev. D44 (1991) 817.
59. A. Bartl, W. Majerotto and B. Mosslacher in Proceedings, e+ e- Collisions at
500 Ge V: The Physics Potential, Munich-Annecy-Hamburg 1991/93, DESY 92-
123A+B,93-123C
60. U. Martyn and G.A. Blair, hep-ph/9910416.
61. S.Y. Choi, A. Djouadi, H. Dreiner, J. Kalinowski and P.M. Zerwas, Eur. Phys. J .
C7 (1999) 123; S.Y. Choi, A. Djouadi, H.S. Song and P.M. Zerwas, Eur. Phys. J.
C8 (1999) 669.
62. T. Tsukamoto, K. Fujii, H. Murayama, M. Yamaguchi and Y. Okada, Phys. Rev.
D51 (1995) 3153.
63. G. Moortgat-Pick, H. Fraas , A. Bartl and W. Majerotto, Eur. Phys. C7 (1999) 113.
64. See C . Heusch , ed, Proceedings, e- e- LC Conference Santa Cruz, Int. J. Mod. Phys.
11 (1996) 1.
65. S. Kraml, Report at the DESY /ECFA Workshop, Oxford 1999.
66. S. Ambrosanio and G.A. Blair, Eur . Phys. J. C in press, hep-ph/ 9905403.
67. N. Arkani-Hamed, S. Dimopoulos and G. Dvali, Phys. Lett. B429 (1998) 263.
68. See e.g. M.E. Peskin, Proceedings, LC Conference, Sitges 1999.
69. G.A. Blair, W. Porod and P.M. Zerwas, DESY 99- 181.
POINT SPLITTING AND U(l) GAUGE INVARIANCE
DIRK OLIVIE
Institute for Theoretical Physics
University of Leuven
B-300l Leuven, Belgium
1. Introduction
It is well known that, in local quantum field theory, one encounters di-
vergences which arise from taking products of field operators at the same
space-time point. As a result, these products do not have a well-defined
meaning. Quite some time ago, Dirac [1] suggested point splitting as a
remedy for this difficulty: instead of taking all the field operators at the
space-time point x , a fixed four-vector E is introduced so that only field
operators with different arguments (x , x± E, . .. ) appear in their products.
As long as E is taken to be different from 0, the products of field operators
are well defined and the theory can be expected to be free of divergences,
i.e., to be regularized. At the end of the calculation, the limit E -+ 0 is
taken, in order to recover the original theory.
A prioiri, there are many ways in which such a point splitting procedure
can be implemented. For gauge theories, however, one must ensure that
the introduction of this new parameter E preserves the invariance under
gauge transformations. It thus appears reasonable that one first attempts
to construct gauge transformations involving products of field operators
taken at different space-time points, which we shall call generalized gauge
transformations. Once such generalized gauge transformations are found,
one can then attempt to construct an action, which is invariant under these
generalized gauge transformations.
In this paper, I shall show how these ideas can be put to work for the
U(1) gauge symmetry, although several aspects can be carried over to the
general Yang-Mills case [2, 3].
95
1.-1. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 95-100.
© 2000 Kiuwer Academic Publishers.
96 DIRK OLIVIE
2. Framework
The standard infinitesimal gauge transformations, 0A, for the photon field
A/1(x) and the Dirac field 1j;(x) in the Abelian U(l) case take the form
OA A/1{x) = -8/1 A{x) ,
6A1j;(X) = -zeA(x)1j;(x) , (1)
6A{J(X) = teA(x){J(x) ,
where A(x) is the gauge parameter. The gauge transformations of the fields
satisfy the U(l) group property: the commutator of two subsequent gauge
transformations vanishes.
In what follows, I shall require that the generalized gauge transforma-
tions preserve this Abelian character, i.e., two such generalized transforma-
tions should commute. This requirement imposes strong restrictions on the
form such transformations can take. Nevertheless, it was found that such
infinitesimal transformations can be constructed and that, for gauge trans-
formations on the fermion fields, they take the form of an infinite series in
powers of the coupling constant e:
=L
00
6A 1j;(x) en 6~n) 1jJ(x). (2)

n=l
This is to be contrasted with the standard U(l) case without point splitting,
where the finite gauge transformations on the fermion field are of infinite
order in the coupling constant, the infinitesimal ones in Eqs. (1) being only
of first order in e.
The action A, invariant under these infinitesimal generalized gauge
transformations, is also an infinite series in the coupling constant e:
A = f d4 x .c(x) = f: en f
n=O
d4 x dn)(x) , (3)
where .c(x) is the Lagrangian density and .c(n)(x) its expansion in powers
of e. The action being an infinite series in e leads to new peculiar Feynman
rules. Besides the one-photon vertex, there are in this generalized theory
also two-, three-, four-, ... photon vertices.
In Sections 3 and 4, I shall present two explicit examples of generalized
gauge transformations.
3. First example
Perhaps the simplest Ansatz one can imagine for generalized gauge trans-
formations is
POINT SPLITTING AND U(l) GAUGE INVARIANCE 97
sr)'IjJ(x) = -zA(x+€)'IjJ(x+2€), (4)

where 10 is the point splitting four-vector as discussed in the introduction.
To satisfy the Abelian group property, one can take, e.g.,
Sr) 'IjJ(x) = - -1 [A(x

2
+ 10) + A(x + 310)] 'IjJ(x + 410)
X+f
l x
+3f dya Aa(Y),
(5)
and so on for the higher order terms.

si
When 10 ~ 0, then 1 ) 'IjJ(x) reduces to the standard gauge transforma-
tion [Eqs. (I)), and Sr) 'IjJ(x) ~ O. The general proof that the higher order
terms sin) 'IjJ(x) indeed exist, can be found in Ref. [4].
The expansion of the corresponding invariant action then yields the
following lowest order results for the Lagrangian density:
dO) = 1{J(x) (qJj 8/-1 - m) 'IjJ(x) -l FJjY(x) FJjY(x), (6)
1{J(x - 10) 1Jj 'IjJ(x + 10) AJj(x), (7)

1 -
= 2" 'IjJ(x - 210) 1 JL 'IjJ(x + 210)
x [[AJL(X - 10) + AJL(x + €)ll~:f dya Aa(Y)
+[l~: dyo Aa(Y) + l~: dya Aa(Y)ll~:f dy~ F/-li3(Y)] . (8)

It is now obvious that through the incorporation of point splitting in the
gauge transformation, one obtains a nonlocal Lagrangian density. It is also
non-Hermitian. If 10 ~ 0 then .c(1) reduces to the standard QED interaction
Lagrangian density, and .c(n) --+ 0 for n :2 2.
The appearance of infinite line integrals is an unattractive feature of

this approach, which will be remedied in Section 4.
4. Second example
In this example, the separation between the different space-time points is
still characterized by a fixed four-vector 10, but for the construction of the
infinitesimal generalized gauge transformations, one takes an average over
the separation in 10 using a weight function p( 1]). The generalized gauge
transformations of the fermion fields are again infinite series, the first order
term being
Sr) 'IjJ(x) = -z [:00 p(a) da [ : p(f3) df3 [:00 pb) d1

x A(x + (a + 1)10) 'IjJ(x + (f3 + 1)10). (9)
98 DIRK OLIVIE
To avoid infinite line integrals, the weight function p( TJ) is taken to be real
and even. It is also normalized
1 +00
-00 dTJ p(TJ) = 1, (10)
which guarantees that or) 'ljJ(x) reduces to the expression in Eqs. (1) when
E -+ O. Finally, it must obey the convolution property
(11)
An example of a function satisfying these four conditions is
p(TJ) = ! sin(TJ) . (12)

7r TJ
In momentum space, the generalized gauge transformations [Eq. (9)]
become:
or) 'ljJ(k) = -z / d4kl / d4k2 o(4)(k - kl - k2) A(kd 'ljJ(k2) p(k 1 . E)

X p(k2· E) P((kl + k2)· E) , (13)
with p(w) the Fourier transform of p(TJ), A{k) and 'ljJ(k) being the Fourier
transforms of A(x) and 'ljJ(x) respectively. The properties of p(TJ) translate
into the following properties of p{w): (i) p(w) is real and even; (ii) p(O) = 1;
(iii) p2(w) = p(w) , implying that p(w) = 0 or 1. For the example in Eq. (12),
one has that
-(w)
P
= {I,
0,
if
if
Iwl < 1
Iwl 2: 1. (14)
To satisfy the Abelian group property of Section 2, one can take, e.g. ,
8~) 'ljJ(k) = -2/ d4kl / d4k2 / d4k30(4)(k - kl - k2 - k3)

A(kd· E
X k1 . E A(k2) 'ljJ(k3) [1 - p((k2 + k3) . E)]
X P(kl . E) p{k2 . E) p(k3 . E) p{{k 1 + k2 + k3) . E) . (15)
and so on for the higher order terms [5]. In spite of the appearance of
kl . E in the denominator in Eq. (15), the expression for o~) 'ljJ (k) is free
of singularities for kl . E -+ O. To see this, it suffices to observe that, for
kl . E -+ 0, p((kl + k2 + k3)· E) -+ p((k2 + k3)· E) and that [1- p(w)] p(w) = 0
for all w.
POINT SPLITTING AND U(l) GAUGE INVARIANCE 99
Through this example, the relation between point splitting and the mit-
igation of ultraviolet divergences becomes apparent. If the function pew) is
taken as in Eq. (14), then, in Eq. (13) for 8~1) 'lj;(k) , the functions P(kl . E) ,
P(k2 . E) and P((kl + k2) . E) cut off the high momentum components in the
direction of E of A(kd, 'lj;(k 2) and 8r) 'lj;(k) respectively. A similar property
holds for the second order term in Eq. (15), and it is also valid for the higher
order terms.
These generalized gauge transformations reduce to the standard gauge
transformations in the limit E ~ 0. Indeed, the first order term 8~1) 'lj;(k)
[Eq. (13)J reduces to
because p(O) = 1. Eq. (16) is exactly the Fourier transform of the standard
°
infinitesimal gauge transformation of Eq. (1). Furhermore, when E ~ 0, the
second order term 8r) 'lj;(k) ~ in Eq. (15) , and similarly for the higher
order terms 8~n) 'lj;(k) for n > 2. Again, one can construct a generalized
action, invariant under these generalized gauge transformations, the details
of which are given in Ref. [5J.
The advantage of the approach in this second example is that it can
also be generalized to the case of Yang-Mills theories [2, 3J. In those cases,
one finds that it is absolutely necessary to average in the generalized gauge
transformations over the different arguments for the field operators.
One is still faced with a shortcoming in this approach: the momentum
cut off of the functions only occures in the direction of the four-vector
E, and, as a result, one cannot expect that all ultraviolet divergences are
regulated.
5. Conclusions
We have shown that it is possible to construct generalized gauge transfor-
mations for which the fields and the gauge parameters are taken at different
space-time points. The separation between the different space-time points
is characterized by a fixed four-vector E. Two examples of generalized gauge
transformations and invariant actions are presented here. It appears that
the generalized infinitesimal gauge transformations of the fermion fields
and the invariant actions are infinite series in the coupling constant e. In
the limit E ~ 0, they reduce to the standard expressions one encounters
in QED. In the second example, the conditions which the weight function
has to satisfy lead to the introduction of a cut off for the large momentum
components of the fields in the direction of E. We are thus led to consider
100 DIRK OLIVIE
generalized gauge transformations which only act on the small momentum

components of the fields.
Acknowledgements
I would like to thank Professors Raymond Gastmans and Tai Tsun Wu for
sharing their insights in this work with me.
References
1. Dirac P.A.M. , Proc. Cambridge Phil. Soc., 30, 150 {1934}
2. Gastmans R., Newton C. and Wu T .T ., Phys. Rev. D, 54, 5302-5314 {1996}
3. Gastmans R. and Wu T .T ., Phys. Rev. D, 57, 1203-1214 (1998)
4. Gastmans R., Newton C. and Wu T .T ., Phys. Lett. B, 341 , 84-89 {1994}
5. Olivie D., Diploma thesis (unpublished), University of Leuven (1997)
NEUTRINO OSCILLATIONS
L. DI LELLA
CERN, CH-1211 Geneva 23
1. Introduction
At present, we know nothing about two basic neutrino properties: their

mass, and whether neutrinos are their own antiparticles or whether 1/ and
j) differ.
In the Standard Model all neutrino masses are set equal to zero 'by
hand'. Under this assumption the neutrino helicity (the spin component
parallel to the momentum) is a good quantum number. If neutrinos are
their own antiparticles (Maiorana neutrinos), the two helicity states are just
the two spin states of the same particle and lepton number is not conserved.
If 1/ and j) differ (Dirac neutrinos), left-handed neutrinos and right-handed
antineutrinos are the only existing physical states; right-handed neutrinos
and left-handed antineutrinos do not exist and lepton number is conserved.
However, if neutrinos are massive the helicity is not a good quantum
number because it depends on the reference frame. Hence massive Dirac
neutrinos and antineutrinos can exist in both helicity states but the inter-
actions of right-handed neutrinos and left-handed antineutrinos with matter
differ from those of left-handed neutrinos and right-handed antineutrinos.
Massive neutrinos could be an important component of hot dark matter
in the Universe. According to Big Bang cosmology, the Universe is filled
with a Fermi gas of neutrinos at a temperature of rv 1.9° K and with a
density of rv 120/ cm3 for each neutrino type.
The neutrino energy density, Pv, normalized to the critical density of
the Universe, Pc is given by
Ov = Pv/Pc = L m v/94 h~ (1)

v
where Pc = 3 HUS7rGN = 1.05 X 104 h~ eV /cm 3 , the neutrino mass mv is

expressed in eV, GN is the Newton constant and Ho = 100 ho km s-1 Mpc- 1
101
J.-J. Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments, 101-169.
102 L.DILELLA
is the Hubble constant. The normalized Hubble expansion rate ho is not

precisely known but is believed to be in the interval 0.6 < ho < 0.8 [1].
Thus a value of L:llmll between 34 and 60 eV would give nil = 1. Present
cosmological models prefer n/l values of the order of 0.2 [2], suggesting that
the heaviest neutrino mass is in the eV range.
2. Direct measurements of neutrino masses

Direct measurements of neutrino masses are very far from being sensitive to
lIf.l and liT mass values of cosmological relevance. All results are consistent
with massless neutrinos. The most recent upper limits are
m(lIe) < 2.5 eV (95% confidence) (2)

from the measurement of the electron energy spectrum from Tritium (3-
decay near the end-point [3];
m(lIf.l) < 0.16 MeV (90% confidence) (3)

from a precise measurement of the J.L+ momentum from 71"+ -t J.L+lIf.l decay
at rest [4];
m(lIT) < 18.2 MeV (95% confidence) (4)
from a study of multi-prong hadronic decays of r± produced in Z -t r+r-
decay [5].
An additional constraint on m(lIe) comes from the search for the neu-
trinoless double-{3 decay process
76G e32 -t 76 Se34 + e- + e- . (5)
This reaction is a second-order weak interaction process in which a
neutrino is emitted by a neutron together with an e- and is reabsorbed
by another neutron in the same nucleus with the emission of another e-.
This process cannot occur for Dirac neutrinos (ll i= ii) because the emitted
neutrino is in fact a ii which cannot be absorbed by a neutron. In the case
of Maiorana neutrinos reaction (5) can only occur if the emitted neutrino
undergoes helicity flip (otherwise it cannot be absorbed by a neutron). The
helicity flip amplitude is proportional to mil/Eli, where Ell is the neutrino
energy, thus reaction (5) can only occur if m/l i= O. For m/l = 0 neutrinoless
double-{3-decay is forbidden also for Maiorana neutrinos.
The most sensitive search so far has been performed by the Heidelberg-
Moscow experiment [6] in the Gran Sasso underground laboratory. No signal
from reaction (5) has been observed, providing a lower bound for the half-
life of reaction (5):
T 1 / 2 > 5.7 X 1025 Y
NEUTRINO OSCILLATIONS 103
at the 90% confidence level. This corresponds to the upper bound

(6)
for Maiorana neutrinos. This bound is modified in the presence of neutrino

mixing (see next Section).
3. Neutrino mixing and oscillations

3.1. THEORY OF NEUTRINO OSCILLATIONS IN VACUUM
Neutrino oscillations are a consequence of the hypothesis of neutrino mixing

first proposed by Pontecorvo [7] and independently by Maki et al. [8]. Ac-
cording to this hypothesis the three known neutrino flavours, V e , v~ and V r ,
are not mass eigenstates but quantum-mechanical superpositions of three
mass eigenstates, Vl, V2 and V3, with mass eigenvalues ml, m2 and m3,
respectively:
(7)
In Eq. (7) Q = e, /-L, T is the flavour index, i = 1, 2, 3 is the index of

the mass eigenstates and U is a unitary 3 x 3 matrix. The relation
also holds, where V = U- l and Via = U~i because U is unitary.

From Eq. (7) it follows that the time evolution of a neutrino with mo-
mentum p produced in the state Va at time t = 0 is given by
(8)
V
where Ei = p2 + mr. If the masses mi are not all equal, the three terms of
the sum in Eq. (8) get out of phase and the state v(t) acquires components
vf3 with {3 =I Q.
The case of two-neutrino mixing is a particularly useful example. In this
case the mixing matrix U is described by only one real parameter 0 (the
mixing angle), and Eqs. (7) and (8) become, respectively,
Va cos 0 Vl + sinO V2
vf3 - sinO Vl + cos 0 V2
and
104 L.DILELLA
The probability to detect a neutrino state v/3 at time t can then be easily
calculated to be
(9)
where we have used the approximation, valid for m <{:: p,
Ei = Jp2 + m~ ~ p (1 + : )
It can be easily demonstrated that, for v(O) = v/3, P/3a(t) is also given
by Eq. (9) . Furthermore, we have
Equation (9) is expressed in natural units. In more familiar units we

can write
Pa /3(L) = sin2(28) sin2 (1.267~;2
L) (10)
where L = ct is the distance from the source in metres, ~m2 = Im~ - mil
is measured in eV2 and E ~ p is the neutrino energy in MeV (the same
equation holds if L is measured in km and E in Ge Y) .
Equation (10) describes an oscillation with amplitude equal to sin2 (28)
and oscillation length A given by
E
A = 2.48 ~m2 (11)
where A is expressed in metres (km), E in MeV (GeY) and ~m2 in ey2.

We note that, if the oscillation length A is much shorter than the size of the
neutrino source or ofthe detector (or of both), the periodic term in Eq. (10)
averages to 1/2 and the oscillation probability becomes independent of L:
Pa /3 = ~ sin2 (28) .
3.2. OSCILLATION EXPERlMENTS
Experiments searching for neutrino oscillations can be subdivided into two

categories:
3.2.1. Disappearance experiments

In these experiments the flux of a given neutrino flavour is measured at a
certain distance L from the source. The presence of neutrino oscillations
has the effect of reducing the flux with respect to the value expected in the
absence of oscillations. The probability measured by these experiments is
Pa:a:(L) =1- L Pa:/3(L)

/3#a:
The sensitivity of these experiments is limited by the systematic uncer-
tainty on the knowledge of the neutrino flux from the source. To reduce
this uncertainty a second detector close to the source is often used in order
to measure directly the neutrino flux.
Disappearance experiments have been performed at nuclear reactors
and at accelerators. The core of a nuclear reactor is an intense source of ve
with an average energy of rv 3 MeV, which can be detected by observing
the reaction ve + p ~ e+ + n. If a ve turns into a vp. or a vr it becomes
invisible because J.L+ or T+ production is energetically forbidden.
Proton accelerators produce vp.'s with energies between rv 30 MeV and
rv 200 GeV. In disappearance experiments the vp. flux is measured by de-
tecting the reaction vp. + nucleon ~ J.L- + hadrons. The energy threshold
for the reaction vp. + n ~ J.L- + P on a neutron at rest is nO.2'MeV.
Disappearance experiments are the only way to detect oscillations in-
volving neutrinos with no coupling to the Wand Z bosons ('sterile' neu-
trinos).
3.2.2. Appearance experiments

These experiments use beams containing predominantly one neutrino
flavour and search for neutrinos of different flavour at a certain distance
from the source.
The sensitivity of these experiments is often limited by the systematic
uncertainty on the knowledge of the beam contamination by other neutrino
flavours at the source. For example, in a typical vp. beam from a high-energy
accelerator the Ve contamination at the source is of the order of 1%.
Searches for Ve and Vr appearance in a beam containing predominantly
vp. have been performed at accelerators. In these experiments the presence
of ve's (vr ) in the beam is detected by observing the reaction Ve (vr ) +
nucleon ~ e- (T-) + hadrons.
106 L. DI LELLA
3.2.3. Parameters of oscillation experiments

Typical parameters of oscillation experiments are listed in Table 1.
TABLE 1. Parameters of oscillation experiments
Neutrino Neutrino Baseline Neutrino Minimum

source type energy accessible
~m2
Sun Ve 1.5 X 1011 m 0.1-18 MeY '" 10- 11 ey2
Cosmic ray vIA Ve '" 10 - 0.2 GeY '" 10- 4 ey2

7r± , I-' ± decay VIA ve 13000 km - several GeY
Nuclear ve '" 20 m few MeY; 10- 1 - 10- 6 ey2
power -300 km (E) '" 3 MeY depending
reactors on baseline
Accelerators VIA Ve '" 20 m '" 30 MeY 10- 3 - 10 ey2
VIA ve -732 km -100 GeY depending
on baseline
3.3. THEORY OF NEUTRlNO OSCILLATION IN MATTER
It was first pointed out by Wolfenstein [9] that neutrino oscillations in

dense matter differ from oscillations in vacuum if ve's are involved. This
effect arises from coherent neutrino scattering at 0° which, in addition to
the Z-boson exchange amplitude (the same for all three neutrino flavours),
in the case of ve's has a contribution from W-boson exchange with the
matter electrons (see the relevant Feynman graphs in Fig. 1).
v v v e-
~
va
~
,
v.
•
~
e,N
:z
I
~
:w
I
>-~-<
e" e-
e,N e- ve
a) b) c)
Figure 1. Feynman graphs for neutrino scattering in matter; a) neutrino-nucleon or

neutrino-electron scattering by Z boson exchange; b) ve-electron and c) ve-electron scat-
tering by W boson exchange.
NEUTRlNO OSCILLATIONS 107
Since scattering at 0° is a coherent process involving an extended target,

the propagation of neutrinos in matter can be described by adding to the
Hamiltonian a potential energy term which for the diagram of Fig. Ib is
given by
In 14 Z
Vw = v2 GFNe ~ 7.63 x 10- A P eV
where G F is the fermi coupling constant, Ne is the number of electrons per
unit volume, p is the matter density in gj cm3 and the ratio Z j A is the
number of electrons per nucleon.
We consider the case of two-neutrino mixing between Ve and vp.:
Ve VIcos BtJ + V2 sin BtJ
vp. = -VI sinBtJ + V2 cosBtJ
where BtJ is the mixing angle in vacuum. We assume that BtJ < 45° and
m2 > ml, where ml(m2) is the VI(V2) mass value. The evolution equation
is
id"iI! =H"iI! (12)
dt
where
is a two-component vector describing the neutrino state at time t and the

Hamiltonian H is a 2 x 2 matrix:
H = Vp2 +M2 +Vz I ~ ~ 1+ Vw I ~ ~ I

where M2 is the square of the mass matrix and Vz is the potential energy
term resulting from Z boson exchange. By using the approximation
/ M2 M2
yp2+M2 ~p+ - ~ E+-
2p 2E
the Hamiltonian can be rewritten as
H = (E + Vz ) I ~ ~ I+ 2E1 M;e+ 2EVw

(13)
M~e
where
1
M;e - 2(JL2 - ~m2 cos 2Bv)
M;p. = M~ = ~ ~m2 sin 2Bv

1
M;p. = 2(JL2 + ~m2 cos 2Bv)
108 L.DILELLA
with J.t2 = mi + m~ and ll.m 2 = m~ - mi. Obviously, the first term of the
Hamiltonian (13) produces no mixing between Ve and Vw
The study of the ideal case of ve's produced in a medium of constant
density is mathematically rather simple and is very useful to understand
the physics of neutrino oscillations in matter. In this case the Hamiltonian
is time-independent and the mass eigenstates can be found by diagonalising
the second matrix in Eq. (13).
The two mass eigenvalues in matter are
and the mixing angle in matter, Om, is given by the equation

11m2 sin 20v
tan 20m = --:--;:;---:--'--:-
ll.m 2 cos 20v - e (14)
where
e= 2VwE ~ 1.526 x 1O- (ZjA)pE eV2
7
(in this equation p is in gjcm3 and the neutrino energy E is in MeV).

The behaviour of the two mass eigenvalues as a function of eis illus-
trated in Fig. 2.
o ~res
e
Figure 2. Neutrino mass eigenvalues in matter as a function of for the case of small
ere.,
mixing angle in vacuum. On the right-hand side of the resonant value, V2 is mostly
Ve while on the left-hand side V2 is mostly vI'"
e
Equation (14) shows that, even if Ov is very small, for = 11m2 cos 20v
the denominator vanishes and the mixing angle in matter, Om, is equal to
45°, which corresponds to maximal mixing. This resonant behaviour was
first noticed by Mikheyev and Smirnov [10] some years after Wolfenstein's
original formulation of the theory of neutrino oscillations in matter. At the
e
resonant value of the difference between the two eigenstates is minimal
and is equal to 8m 2 sin 28v .
The oscillation length in matter, Am, is longer than in vacuum and is
given by
where Av is the oscillation length in vacuum given by Eq. (11). The maxi-
mum value of Am is reached at resonance, where Am = Avl sin 28v.
The potential energy term Vw changes sign for ve (see the Feynman
graph of Fig. lc). As a consequence, the difference between the two mass
eigenvalues increases monotonically with density. There is no resonance,
therefore, in the case of antineutrinos.
4. Solar neutrinos
4.1. STANDARD SOLAR MODEL
As all visible stars, the Sun was formed from the gravitational collapse
of a cloud of gas consisting mostly of hydrogen and helium. This collapse
produced an increase of the core density and temperature resulting in the
ignition of nuclear fusion reactions. A state of hydrostatic equilibrium was
reached when the kinetic and radiation pressure balanced the gravitational
forces preventing any further collapse.
There are several nuclear fusion reactions occurring in the Sun core, all
having the effect of transforming four protons into a He4 nucleus:
4p -+ He 4 + 2e+ + 2l1e
This reaction is followed by the annihilation of the two positrons with
two electrons, so the average energy produced by this reaction and emitted
by the Sun in the form of electromagnetic radiation is
Q = (4mp - MHe + 2me)c2- < E(2l1e) >~ 26.1 MeV
where m p , MHe, me are the proton, He4 nucleus and electron mass, re-
spectively, and < E(2l1e) >~ 0.59 MeV is the average energy carried by
the two neutrinos. The Sun luminosity is measured to be [1]
Lo = 3.846 X 1026 W = 2.400 X 1039 MeV Is
From the values of Q and Lo it is possible to calculate the rate of lie
emission from the Sun:
110 L. DI LELLA
from which one can calculate the solar neutrino flux on Earth using the
average distance between the Sun and the Earth (1.496 x lO11 m):
~II ~ 6.4 X 1010 cm- 2 S-l
The Standard Solar Model (SSM), which has been developed and con-
tinuously updated by J.N. Bahcall during the past 20 years [11],[12] predicts
the energy spectrum of the solar neutrinos. The main assumptions of the
SSM are:
(i) hydrostatic equilibrium;
(ii) energy production by fusion;
(iii) thermal equilibrium (i.e., the thermal energy production rate is equal
to the luminosity);
(iv) the energy transport inside the Sun is dominated by radiation.
Table 2 shows a list of Sun parameters.
TABLE 2. Sun parameters
Luminosity 3.846 x 10 26 W
Radius 6.96 x 10 5 Km
Mass 1.989 x 1030 Kg
Core temperature Tc 15.6 x 106 oK
Surface temperature T. 5773 oK
Hydrogen content in the core (by mass) 34.1%

Helium content in the core (by mass) 63.9%
The age of the Sun (4.6 X 109 years) is also known. The SSM calcula-
tions are performed by adjusting the initial parameters, by evolving them
to the present day and by comparing the predicted and measured properties
of the Sun. The initial composition of the Sun is taken to be equal to the
present day measurement of the surface abundances. If the predicted prop-
erties disagree with the measured ones, the calculations are repeated with
different initial parameters until agreement is found. These calculations
require the knowledge of the absolute cross-sections for nuclear reactions
in a very low energy region where little information is directly available
from laboratory experiments. Another important ingredient in these cal-
culations is the knowledge of the opacity versus radius which controls the
energy transport inside the Sun and the internal temperature distribution.
There are two main nuclear reaction cycles in the Sun core:
(i) The pp cycle, responsible for 98.5% of the Sun luminosity. This cycle
involves the following reactions:
p+p -+ e+ + Ve + d (15)
p+d -+ I +He3 (16)

He3 + He3 -+ He 4 + p + p (17)
where the second He3 nucleus in the initial state of reaction (17) is
produced by another sequence of reactions (15) and (16).
Reactions (15) through (17) represent 85% of the pp cycle. In the
remaining 15% reaction (17) is replaced by the following sequence of
reactions:
He3 + He4 -+ T + Be7 (18)
e- + Be 7 -+ Ve + Li7 (19)
p + Li7 -+ He 4 + He4 (20)
In approximately 1.9 x 10- 3 of the cases reactions (19) and (20) are
replaced by
p + Be7 -+ T + B8 (21)
B8 -+ Be8 + e+ + ve (22)
Be8 -+ He4 + He4 (23)
Reaction (15) is replaced in 0.4% of the cases by the three-body fusion
reaction
p + e- + p -+ d + ve (24)
Finally, in an even smaller fraction of the cases (I'V 2.4 X 10- 5 ), reaction
(17) is replaced by
(25)
It can be seen that in the pp cycle ve's are produced by the five reac-
tions (15), (19), (22), (24) and (25). These neutrinos will be denoted
as v pp , VBe, VB, vpep and vhep, respectively. While v PP ' VB and vhep
have a continuous energy spectrum, VBe and vpep are mono-energetic
because they are produced in two-body final states.
(ii) The eNO cycle, which involves heavier elements. This cycle consists
of the following chains of reactions:
p+N 15 -+ C 12 +He4
p + C 12 -+ ,+ N 13
112 L.DILELLA
N 13 --t e+ + Ve + C13 (26)

P + C 13 --t 'Y + N 14
P + N 14 --t 'Y + 0 15 (27)
0 15 --t e+ + Ve + N I5 (28)
and
p + N 15 --t 'Y + 0 16
P + 0 16 --t 'Y + FI7
F17 --t e+ + Ve + 0 17 (29)
P + 0 17 --t N 14 + He4
followed by reactions (27) and (28). As for the pp cycle, the two chains
of reactions in the CNO cycle have the overall effect of transforming
four protons into a He4 nucleus. Production of Ve occurs in reactions
(26), (28) and (29). These neutrinos will be denoted as VN, Va and
VF, respectively.
Figure 3 shows the Ve flux as a function of energy, as predicted by the

SSM for the different reactions. The vpp flux is the dominant component.
However, neutrino cross-sections increase rapidly with energy (typically as
E; for energies well above threshold), so these neutrinos are not among the
easiest ones to detect because of their low-energy. Figure 3 displays also
the energy threshold for the capture reaction
Ve + (A,Z) --t e- + (A,Z + 1)

for a variety of nuclear isotopes. The SSM also makes predictions on neu-
trino production as a function of radius.
It must be finally pointed out that, while solar neutrinos arrive on Earth
approximately 500 s after being produced, it takes of the order of 106 years
for the energy produced in the same reactions to be transported from the
Sun core to its surface. Thus the Sun luminosity which is measured at
present is associated with neutrinos which reached the Earth rv 106 years
ago. This is not considered to be a problem for the SSM because the Sun is
a star on the main sequence, with no appreciable change of properties over
rv 108 years.
Solar Neutrino
Spectrum
0.1 10
NE[JfRINO ENERGY (MeV)
Figure 3. Solar neutrino energy spectrum as calculated from the SSM [l1J. Energy
thresholds for various neutrino detection processes are shown on top.
4.2. SOLAR NEUTRINO EXPERIMENTS
4.2.1. The Homestake experiment

Solar neutrinos were successfully detected for the first time in an experiment
performed by Davis and collaborators [13] in the Homestake gold mine
(South Dakota, U.S.A.). The method consists in measuring the production
rate of 37 A from the capture reaction
(30)
which occurs in a 390 m2 tank filled with 615 tonnes of perchloroethylene
(C2CI4, a commonly used cleaning fluid). The isotope 37 CI represents 24%
of all natural chlorine, so there are approximately 125 tonnes of 37 CI in the
tank. The neutrino energy threshold for this reaction is 0.814 MeV, so this
reaction is not sensitive to the vpp component (see Fig. 3).
The tank is installed deep underground in order to reduce 37 A produc-
tion by cosmic rays. Every few months, Argon is extracted from the tank
by N2 flow. It is then separated, purified, mixed with natural Argon and
used to fill a proportional counter. The presence of 37A in the counter is
then detected by observing its decay which occurs by electron capture with
a half-life time 71/2 = 34d:
e- + 37A -+ Ve + 37CI .
In this reaction an X-ray or an Auger electron emitted from the atomic
transition to the orbital state left empty after electron capture is detected in
114 L.DILELLA
the proportional counter. The extraction efficiency is measured by injecting

a known amount of 37A in the tank. On average, the 37A production rate is
of the order of 0.5 atoms/day.
It has become customary to express the solar neutrino capture rate in
Solar Neutrino Units or SNU (1 SNU corresponds to 1 capture/s from 1036
nuclei). The result of the Homestake experiment averaged over more than
20 years of data taking, is [14]
Rexp(37C) = 2.56 ± 0.16 ± 0.16 (31)
where the first error is statistical and the second one represents the sys-
tematic uncertainties.
Table 3 shows the SSM predictions, as calculated by Bahcall et al. [12].
The total rate is predicted to be Rt h( 37 CI) = 7.7~t~ SNU, which dis-
agrees with the measured value. An independent SSM calculation by Turck-
Chieze et al. [15] predicts Rth (37 CI) = 6.4 ± 1.4 SNU which is again larger
than the measured value.
TABLE 3. Solar neutrino contributions to reaction (30) as

predicted by the SSM [12]
Solar neutrino component 37 A production rate (SNU)

VB 5.9
VBe 1.1
va 0.4
vpep 0.2
VN 0.1
Total 7.7::~:~ ± 3.0
4.2.2. Gallium experiments

Two experiments, GALLEX and SAGE, have measured the rate of the
reaction
(32)
which has a neutrino energy threshold of 0.233 MeV and is sensitive, there-
fore, to the vpp contribution.
GALLEX, installed in the Gran Sasso underground laboratory (Italy),
uses 30.3 tons of Gallium dissolved in HCl. SAGE, installed in the Baksan
underground laboratory (Russia) uses 57 tons of metallic Gallium which is
liquid at 40°C. The fraction of 71Ga isotope in natural Gallium is 39.7%.
In both experiments 71Ge is extracted every 3-4 weeks by means of

physical and chemical methods and converted to GeH4. This is a gaseous
substance which is used to fill a proportional counter built from special, low
radioactivity materials. This counter is used to detect the atomic X-rays
emitted from 71Ge decay which occurs by the electron capture reaction
(33)
with a half-life of 11.43 d. The X-ray time and energy distributions provide
evidence for 71Ge production by solar neutrinos.
Both experiments have performed convincing checks of the 71Ge extrac-
tion efficiency, which include the use of a very intense 51Cr source producing
mono-energetic neutrinos from the electron capture reaction
with a half-life of 27.7 d. The neutrino energy from this source is 0.75 MeV
and the initial flux at the detector corresponds to several times the solar
neutrino flux.
In the GALLEX experiment the 71Ge extraction efficiency has been
measured directly by injecting known quantities of 71 As in the tank. This
isotope decays to 71Ge with a half-life of 2.72 d, either by {3+ decay or by
electron capture.
The 71Ge production rate from solar neutrinos is measured to be
R( 71 Ga) = 77.5 ± 6.2~g SNU
by the GALLEX experiment [16]; and
R( 71 Ga) = 66.6+-7.1-4.0
6 .8+3.8 SNU
by SAGE [17]. In both experiments the first error is statistical and the
second one represents the systematic uncertainties.
After adding in quadrature the statistical and systematic errors, the
weighted average of the two results is
R( 71 Ga) = 72.7 ± 5.7 SNU . (34)
The SSM predictions are shown in Table 4. Again, the measured 71Ge
production rate is much lower than the SSM predictions.
116 L. DI LELLA
TABLE 4. Solar neutrino contributions to reaction (32), as

predicted by the SSM
Solar neutrino component 7lGe production rate (SNU)

Ref. [12] Ref. [15]
Vpp 69.6 70.6

Vpep 2.8 2.8
VBe 34.4 30.6
VB 12.4 9.3
VN 3.7 3.9
Vo 6.0 6.5
Total 129~: 124±5
4.2.3. Super-Kamiokande
Super-Kamiokande is a real-time experiment which uses an underground
detector installed in the Kamioka mine 350 km west of Tokyo.
The inner detector consists of a cylindrical tank filled with 32,000 tons
of water. Approximately 40% of the tank surface are covered by 11,146
photomultipliers with a diameter of 50 cm and pointing towards the liquid.
The inner detector is surrounded by an additional layer of water, with
a thickness of 2 m and seen by 1,881 photomultipliers with a diameter of
20 cm. This outer detector is used to identify charged particles entering the
detector from outside.
The inner detector is used as an imaging Cerenkov counter. Charged
particles with v / c ~ 1 produce Cerenkov light at an angle of ,...., 41 0 to
their direction of flight and the pattern of hit photomultipliers and their
relative timing provide information on the particle direction and origin in
the detector volume.
Solar neutrinos are detected by the scattering reaction
(35)
which is suppressed by approximately 1/6 for vIJ and Vr with respect to

Ve. The electron energy scale is calibrated by sending an electron beam of
5-16 MeV energy from a LINAC into the inner detector at various depths.
The threshold for solar neutrino studies is as low as 5.5 MeV. Only events
produced at a distance of at least 2 m from the inner detector walls are
considered. This fiducial volume contains 22,500 tons of water.
The detected electron from reaction (35) has a very strong directional
correlation with the incident neutrino. This property is used to demonstrate
the solar origin of the events, as shown in Fig. 4 which displays the dis-
tribution of the angle between the electron direction and the Sun-to-Earth
direction at the time of the event. The peak at cos 8 sun = 1 is due to solar
neutrinos.
The Super-Kamiokande experiment began data taking in May 1996 and
has reported results from a run of 708 days [18]. The solar neutrino flux
with an electron energy threshold of 6.5 MeV is measured to be
<)1/ = (2.44 ± 0.04 ± 0.07) x 106 cm- 2 S-1 (36)
c 0.3 r----~------~---........,
e80.25 - - Best-fit Super-Kamiokande 297.4day
~
CO 0.2
~
~0.15
>
w
0.05
-0.5 o 0.5 1
COS9sun
Figure 4. Distribution of the cosine of the angle between the electron direction and the
Sun-to-Earth direction, as measured by Super-Kamiokande. The curve is a best fit to the
data.
where the first error is statistical and the second one represents system-
atic uncertainties. This value is 47% of the SSM prediction [12], <)SSM =
(5.15~g:~~) x 106 cm- 2 S-1. Obviously with a detection threshold of 6.5
MeV the experiment is only sensitive to VB (see Fig. 3).
4.3. INTERPRETATION OF THE SOLAR NEUTRINO DATA
The measured solar neutrino event rates are dominated by three compo-
nents of the JrP cycle: vPP ' //Be and //B.
We parametrize the deviations of the measured fluxes from the SSM
predictions as follows:
Xpp <Pm (vpp ) / <PSSM (vpp );
XBe <Pm (//Be) / <PSSM (//Be ) j
XB <Pm (va)/<PSSM (va),
118 L.DILELLA
where by <Pm we denote measured fluxes and by <PSSM the predictions of

Ref. [12] . With these parameters the result of the Homestake experiment
(see Eq. 31) can be written as
1.2 XBe + 5.9xB + 0.6 = 2.56 ± 0.23 SNU (37)
where the left term contains the SSM predictions for variable lIBe and lIB
fluxes and the additional contribution (0.6 SNU) from all other solar neu-
trino components (see Table 3), while the right term is the experimental
result with statistical and systematic uncertainties added in quadrature.
Similarly, with the help of Table 4 the combined result of the two Gal-
lium experiments can be written as
69.6 xpp + 34.4 XBe + 12.4 XB + 12.6 = 72.7 ± 5.7 SNU . (38)
Finally, Super-Kamiokande measures directly XB:
XB = 0.47 ± 0.02 . (39)
The solution of the three equations is
Xpp = 1.03 ± 0.14 ; XBe = -0.68 ± 0.22 .

If the contributions from all other neutrino components are set equal to
zero in Eqs. (37, 38), the solution becomes
Xpp = 1.04 ± 0.14 ; XBe = -0.18 ± 0.22 .

In both cases XBe is consistent with zero.
The apparent absence of lIBe in the solar neutrinos is a real puzzle,
because neutrinos from B8(lIB) are observed by Super-Kamiokande (albeit
at a rate which is 47% of the predicted one) and, as explained in Section
5.1, B8 is formed from the fusion reaction
which implies that Be7 must exist in the Sun. This in turn implies the
occurence of the reaction
which is responsible for lIBe production, at a rate three orders of magnitude

higher than the reaction responsible for B8 formation which is strongly
suppressed by Coulomb repulsion.
There are three possible explanations to this puzzle:
(i) At least two of the three measurements of the solar neutrino flux are
wrong;
(ii) There is a basic flaw in the SSM, resulting in unreliable predictions of
the solar neutrino flux (however, the SSM correctly predicts the results
of helioseismological observations [19] which depend on the tempera-
ture profile of the Sun);
(iii) The lIBe'S are produced as Ve in the core of the Sun but are no longer
Ve when they reach the Earth.
This last explanation, which we assume to be the correct one, implies the
occurence of neutrino oscillations.
4.4. VACUUM OSCILLATION SOLUTIONS
The experimental results on solar neutrinos can be explained by oscillation

parameters which strongly suppress the lIBe component (Ev = 0.861 MeV)
when these neutrinos reach the Earth at a distance L ~ 1.5 X 1011 m
from the Sun. These are the so-called vacuum oscillation solutions, also
nick-named 'just so' solutions because they require a precise mathematical
relation between three unrelated physical quantities (dm2, L and the lIBe
energy).
These solutions, as derived by the Super-Kamiokande collaboration [18],
are shown in Fig. 5a. The dm2 values are in the range 4x 10- 11 - 5 X
10-10 eV2 and the mixing angle is large (sin229 > 0.6).
Two measurable effects are expected from these solutions:
(i) A seasonal variation of the measured solar neutrino flux exceeding the
6.7% solid angle variation associated with the excentricity of the Earth
orbit around the Sun. The statistical precision of the present Super-
Kamiokande results is insufficient to reach a definitive conclusion on
this point (see Fig. 6a);
(ii) An energy dependent suppression of the solar neutrino flux. A signif-
icant energy dependence is indeed observed in the Super-Kamiokande
experiment for electron energies above 13 MeV, in agreement with a
vacuum oscillation solution (see Fig. 6b). However, it has been recently
pointed out that the SSM prediction of the l1:tep component of the so-
lar neutrino spectrum (see Fig. 3) is affected by very large theoretical
uncertainties resulting from uncertainties on the cross-section for reac-
tion (25). Hence one cannot exclude that the excess of events observed
at high energies in the distribution of Fig. 6b could be due to a Vhep
flux exceeding the SSM prediction by an order of magnitude or more.
120 L. DI LELLA
99% Confidence Level
-
c:.-
~ -9.4
a)
'-"
-10
-10.2
-10.4
-10.60 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 2 1
sin 2e
-~ ~---------------------------. b)
~~
~9"k C.l. III
95%C.l. III
-5
-6
-7
-8 .......'-J.,..;...L..J....'-'-I..L...l...J-Ji.J....L...L..I-....w..............o..J....Ju....L..L.L'-J.,..;...J....I....I...J-J~
-4 ~ .5 ~ -2.5 -2 -1.5 -1 -0.5 2 0
log (sin 29)
Figure 5. Vacuum oscillation (a) and MSW (b) solutions to the solar neutrino results
under the assumption of two-neutrino mixing [18].
~ 1 .----.--,-----r-...,.--r--;--.-----,r---.-~__r-.., a)
~ SK 708day 6.5-20MeV 22.5kt
~ Sf< 708day 11.5-20MeV 22.5kt
~ w/o eccentricity correction
caO.8 (Preliminary)
C5
o
·r. . . ·
,
0.6 1......
...
1
T
+
~~-f~_
0.4 I
0.2
"' 1
Ir SK SLE 419day + LE 70Bday 22.5kt ALL
b)
<D
5.5-20MeV (Preliminary)
~ 0.9
! stat error
~
iii O.B ! Istat 2+ syst 2
Cl
0.7 SLE ~ LE
0.6
0.5
0.4
0.3
0.2
0.1
O~~~~~--------~----~--~~--~~
5 7.5 10 12.5 15
Energy(MeV)
Figure 6. a) Yearly variation of the measured solar neutrino signal normalized to the
average predicted signal. The curve represents the variation expected from the excentric-
ity of the Earth orbit around the Sun; b) Ratio between measured and predicted solar
spectrum. The curves show the expectations for the vacuum and MSW (small mixing
angle) solution.
122 L.DILELLA
4.5. MATTER ENHANCED SOLUTIONS
Another class of oscillation solutions to the solar neutrino problem can

be found in the framework of the theory of neutrino oscillations in mat-
ter described in Section 3.3. These are the so-called Mikheyev-Smirnov-
Wolfenstein (MSW) solutions.
For neutrinos propagating through the Sun, the density p varies along
the trajectory from a value higher than 100 gjcm3 in the core to much
less than 1 gjcm3 in the outermost layers. The ratio Zj A also varies across
the Sun because of the varying hydrogen abundance. Hence, in Eq. (12)
the Hamiltonian depends on time. For a given set of mixing parameters
ml, m2 and Ov, Eq. (12) can be solved numerically with the condition that
the initial neutrino state is a pure Ve state, using the SSM prediction for
the solar density, Zj A ratio and distribution of neutrino origins inside the
Sun core.
The ideal case of constant density discussed in Section 3.3 represents
a good approximation to a class of solutions of relatively short oscillation
length for which the variation of the solar density over an oscillation length
is negligible (the so-called adiabatic solutions):
1 dp
--Am ~ 1
pdr
where r is the distance from the Sun centre. For such solutions the neutrino
can be described as a superposition of mass eigenstates with slowly varying
eigenvalues and mixing angle. In this case, if for a Ve at production the
e
condition > ~m2 cos 20v is satisfied, then Om is larger than 45° (see Eq. 14)
and the dominant mass eigenstate is 112. If, furthermore, the adiabaticity
condition is satisfied also at resonance, where Am is maximal, then the V2 --t
Vl transition probability is negligible and the dominant mass eigenstate is
still V2 when the neutrino emerges from the Sun. However, the V2 eigenstate
in vacuum is mostly vlJ because Ov < 45°. Thus the Mikheyev-Smirnov
resonance offers an elegant way to explain the solar neutrino problem even
if the mixing angle in vacuum is small.
It must be pointed out that, in the case of small mixing angle, only the
e
ve's produced with > ~m2 cos 20v may emerge from the Sun as vlJ's as
e
a result of the MSW effect. As depends linearly on the neutrino energy
E, this condition is satisfied only by neutrinos produced above a critical
energy which depends on the mixing parameters.
The results from the latest analysis of the solar neutrino event rates in
terms of matter enhanced oscillations [18] are shown in Fig. 5b. For each
experiment the measured event rate corresponds to a region of allowed pa-
rameters in the sin2(20v ), ~m2 plane. This region consists of a vertical band
at large mixing angles, of a horizontal band at constant ~m2 corresponding

to adiabatic solutions and extending to small mixing angles, and of another
band merging into the two previous ones for which the allowed values of sin
28v decrease with increasing ~m2 (see Ref. [20]). Since the processes used
to detect solar neutrinos have different energy thresholds, these regions do
not coincide and the oscillation parameters which describe all available data
are defined by their overlap. As shown in Fig. 5b there are three possible
solutions, two with large mixing angles and ~m2 ~ 10- 7 or 4 xl0- 5 eV 2,
and one with small mixing (sin2 28 ~ 0.005) and ~m2 ~ 6 x 10- 6 eV2.
Two effects are expected from MSW solutions:
(i) a possible day-night variation of the solar neutrino event rate, with
an increased rate at nights from matter enhanced oscillations for neu-
trinos crossing the Earth, resulting in an increase of the Ve flux. The
most recent measurement of the day-night asymmetry in the Super-
Kamiokande experiment gives [21]
Rd-Rn
Rd + Rn = -0.065 ± 0.031 (stat.) ± 0.013 (syst.) ,
where Rd (Rn) is the day (night) event rate.

(ii) an energy dependent suppression of the electron energy spectrum (see
Fig. 6b).
4.6. FUTURE SOLAR NEUTRINO EXPERIMENTS
4.6.1. SNO
The Sudbury Neutrino Observatory (SNO) is a solar neutrino detector in-
stalled in the Creighton mine near Sudbury, Ontario, at a depth of 2070 m
(5900 m water equivalent) [22]. The detector consists of a spherical acrylic
vessel with a 6 m radius containing'" 1000 tons of high purity heavy water
(D20) surrounded by 7800 tons of ultra-pure water for shielding purposes.
Cerenkov light produced in the heavy water is collected by 9456 photomul-
tipliers with a diameter of 20 cm located on a concentric spherical surface
at a radius of 9.5 m.
As for the SuperKamiokande experiment, solar neutrinos are detected
by observing the elastic scattering reaction ve- -+ ve-, which is dominated
by the Ve component of the solar flux and provides precise information on
the incident neutrino direction. With an electron energy threshold set at
5 MeV the expected event rate is 1.4 d- 1 assuming a 50% reduction of the
Ve flux. However, in heavy water the charged current reaction
124 L.DILELLA
also occurs, at a rate of 12.7 events/d for a 5 MeV threshold and for the
same reduction of the Ve :flux. A measurable asymmetry with respect to the
Sun position in the sky is also present in this reaction.
The main feature of SNO is its anticipated capability to detect the
reaction
v+d-+p+n+v (40)
which has the same cross-section for all three neutrino :flavours and mea-
sures the total solar neutrino :flux. Any significant difference between the
neutrino:flux measured from reaction (40) and that measured from charged-
current reactions would provide, therefore, unambiguous proof of neutrino
oscillations.
Two methods are used to detect reaction (40):
(i) MgCl2 is added to the heavy water. In this case neutrons from reaction
(40) undergo the capture process n + 35Cl-+ 36 Cl+, and the 8.5 MeV
,-ray converts in the heavy water.
(ii) Re3 proportional counters are inserted in the heavy water volume and
detect the monlrenergetic signal from the capture process n + Re 3 -+
R3+ p.
The event rate from reaction (40) is expected to be 5.5/d for a neutron
detection efficiency of 40%.
The SNO detector is presently completely filled with heavy water and
data taking has recently started. After one year, it is planned to add MgCb
and to take data for another year. Then, after removal of MgCb data taking
will continue with an array of Re3 proportional counters in the heavy water
volume.
4.6.2. Borexino
Borexino is an experiment presently under construction at the Gran Sasso
National Laboratory [23]. The detector consists of a spherical acrylic vessel
of 8.5 m diameter filled with very high purity, low activity liquid scintillator
and viewed by an array of 1650 photomultipliers located on its surface.
The relative timings of the photomultiplier signals provides information
on the event position within the detector volume: only events occurring
within a central fiducial volume corresponding to 100 tons of scintillator
are considered. The entire detector is immersed in a cylindrical tank 16.5
m high with a 16.5 m diameter filled with high purity water and acting as
a shield.
The aim of the experiment is to detect v - e- elastic scattering with an
energy threshold as low as 0.25 MeV. IT this is achieved, the experiment is
sensitive to the lISe component of the solar:flux (E = 0.861 MeV) which is
expected to be strongly suppressed if neutrino oscillations are indeed the

solution of the solar neutrino problem (see Section 4.3).
Figure 7 shows the expected electron energy distribution. The recoil
electrons from "'Be elastic scattering on e- at rest have a fiat kinetic energy
spectrum with the end-point at 0.664 MeV. In the absence of oscillation,
and with a threshold as low as 0.25 MeV, one expects a contribution of
50 events/d from these neutrinos. Because of the scintillation light isotropy,
the identification of the solar origin of the signal relies on the observation
of its seasonal variation associated with the excentricity of the Earth orbit
around the Sun.
Data taking is expected to start in the year 2000.
10000
iii
~
1000
>co
-
~
N
0
ci
100
J!l
c:
::J
0
t.l
10
o 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5

Energy MeV
Figure 7. Simulated spectra of the signal expected in Borexino from the Standard Solar
Model (upper line) and from the standard flux modulated by the neutrino oscillation ex-
pected from the MSW (small mixing angle) solution. The dotted line shows the expected
background.
4.6.3. KAMLAND
The Kamioka Liquid Scintillator Anti Neutrino Detector (KAMLAND) is
not a solar neutrino experiment. It is discussed here because it is sensitive
to oscillation parameters which could explain the solar neutrino problem
[24].
The detector is a transparent sphere with a diameter of 13 m filled with
scintillating isoparaffin oil. This sphere is itself contained in a larger, con-
centric sphere (18 m diameter) filled with pure isoparaffin oil. Scintillation
light from the inner sphere is collected by 1300 photomultipliers located
126 L. DI LELLA
on the surface of the outer sphere. The entire system is immersed in high
purity water and installed in the Kamioka mine at a depth of 2700 m of
water equivalent.
KAMLAND aims at detecting the ve produced by five nuclear reactors
located at distances between 150 and 210 km from the detector and pro-
ducing a total thermal power of 127 GW. The ve, with an average energy
of 3 MeY, are detected by measuring the e+ signal from the reaction
(41)
followed by the late, signal from the neutron capture reaction np -+ d, (E-y
= 2.2 MeY) which occurs after neutron thermalization.
When all reactors run at full power, the event rate is expected to be
3 d- 1 with a signal-to-noise ratio of 10. The background is measured by
observing the variation of the event rate with the reactor power.
Because of its large distance from the reactors and of the low ve energy,
KAMLAND is sensitive to !l.m2 > 7 x 10-6 ey2 and sin229 > 0.1, a re-
gion which includes the large mixing angle, large !l.m 2 MSW solution (see
Fig.5b).
KAMLAND will begin data taking in the year 2000.
5. Atmospheric neutrinos
5.1. ORIGIN OF ATMOSPHERIC NEUTRINOS
Since the total thickness of the atmosphere is I"V 103 g/cm2 , which is
equivalent to 10 interaction lengths, the interaction of a primary cosmic
I"V
ray in the upper layers of the atmosphere results in the development of a

hadronic shower leading to a flux of neutrinos from charged pion and muon
decay. These neutrinos have energies ranging from 0.1 GeY to several
I"V
GeY. Their interaction rate is of the order of 100ly for a target mass of
1000 tons.
Since a vI' is produced from both 1r± and J1.± decay, and a Ve from J1.±
decay only, one expects the ratio between the vI' and Ve fluxes on Earth to
be of the order of 2 if both 1r± and J1.± decay in the atmosphere. This is
indeed a very good approximation for neutrinos with energies lower than
3 GeY. At higher energies, this ratio increases because the J1.± decay path
increases with energy and the fraction of J1.± decaying in the atmosphere
and producing Ve or ve decreases.
Calculations of atmospheric neutrino fluxes [25] are affected by sizeable
uncertainties which result from uncertainties on the composition and energy
spectrum of the primary cosmic rays and on secondary particle distribu-
tion. In addition, these calculations ignore for simplicity the lateral shower
development and treat the problem in one dimension only. The final un-
certainty affecting the v ll and Ve fluxes on Earth is estimated to be of the
order of ± 30%. However, because of partial cancellations, the uncertainty
on the predicted vll/ve ratio is believed to be of the order of ± 5%.
5.2. Vp./ve FLUX RATIO MEASUREMENTS
Six underground experiments have measured the atmospheric neutrino fluxes

by detecting quasi-elastic interactions:
V Il (ve )+ n ~ J.L-(e-) + p
iill(iie ) + P ~ J.L+(e+) + n .
Three experiments (Kamiokande [26], IMB-3 [27] and Super-Kamio-
kande [28]) detect the Cerenkov light ring produced by relativistic particles
in water. The other three experiments, FREJUS [29], NUS EX [30] and
Soudan-2 [31] use calorimeters with high longitudinal and transverse seg-
mentation. FREJUS and NUSEX took data in the 80's, while Soudan-2 is
presently running.
Muons from quasi-elastic reactions appear in all these detectors as sin-
gle, penetrating tracks. IT the muon stops in the detector and decays, the
decay electron can also be observed. The Super-Kamiokande experiment
accepts also muon events from deep-inelastic neutrino interactions in the
water if the muon exits the inner detector. These 'multi-ring' partially con-
tained events are produced by vll or iill with an average energy of 15 GeV.
Electrons from quasi-elastic reactions produce single electromagnetic
showers consisting of many short tracks which are easily identified in the
calorimeters and result in diffuse Cerenkov light rings in the water detectors.
For most experiments the electron-muon identification capabilities have
been measured using test beams from accelerators. For the Kamiokande
and Super-Kamiokande experiments the mis-identification probability is
measured to be less than 2% [32].
The comparison between the measured and predicted vll/ve ratio for
the six experiments is shown in Table 5 which lists the values of the double
ratio R defined as
R = (vll/ve) measured
(vl'/ve) predicted .
With the exception of the values measured by FREJUS and NUSEX, all
values of R are significantly lower than the expectation (R = 1). The NU-
SEX result is affected by a large statistical error, while the FREJUS result
is not confirmed by Soudan-2 which uses the same detection technique. It
can be concluded, therefore, that a small R value (of the order of 0.6) is
now firmly established.
128 L. DI LELLA
TABLE 5. Measured values of R. The momentum range of observed charged

leptons (Kamiokande, 1MB, SuperKamiokande) or the visible energy range
(Frejus, NUSEX, Soudan 2) is denoted by PI.
Experiments Pl(MeV Ic) Exposure R Ref.

(kt-yr)
Kamiokande e:100-1330 7.7 0.60±8:8g ± 0.05 [26]
JL:200-1400
e:1330- 6.0",,8.2 0.57±8:8~ ± 0 0. 7 [26]
JL:1400-
1MB e:l00-1500 7.7 0.54±0.05±0.12 [27]
JL:300-1500
Frejus e:200- 2.0 1.00±0.15±0.08 [29]
JL:200-
NUSEX e:200- 0.74 0.99±8: ~g [30]
JL:200-
Soudan 2 e:150- 2.83 0.58±0.11±0.05 [31]
JL:100-
Super-Kamiokande e:100-1330 45.4 0.668±g:8~~±0.052 [28]
JL:200-1400
e:1330- 45.4 0.663±8:8!f ± 0.079 [28]
JL:1400-
5.3. ZENITH ANGLE DISTRIBUTION
The flight path of atmospheric neutrinos from the production point to the
detector, L, varies enormously with the zenith angle Bz. For example, neu-
trinos impinging on the detector from above (cosBz = 1) are produced
I"V10 km above the detector, while upward going neutrinos (cosB z = -1) have
traversed the Earth and so have travelled for 13000 km before reaching
I"V
the detector. Also, the higher the neutrino energy, the better the outgoing
charged lepton follows the incident neutrino direction. Hence the charged
lepton zenith angle is a direct measurement of L. All atmospheric neutrino
detectors can be seen, therefore, as disappearance experiments with vari-
able neutrino energies and path lengths. Measurements of the zenith angle
distributions are a sensitive way to search for neutrino oscillations.
The Kamiokande experiment [33] published a dependence of R on Bz
which disagreed from the expected shape at the level of 2 standard de- I"V
viations. The Super-Kamiokande ()z distributions [28], with a much larger

event sample, are shown in Fig. 8. It is clear that there are less muon
events in the upward direction (negative cos ()z) than expected, while the
number of downward going muons is consistent with the expectation. For
electrons, however, the distributions agree with expectations. The distri-
butions dR/dcos ()z, also shown in Fig. 8, reflect the distortion seen in the
muon distributions.
Figure 9 shows the up-down asymmetry defined as (U-D)/(U+D), where
U(D) is the total number of events with cos ()z < -0.2 (cos()z > 0.2), as
a function of the charged lepton momentum. While for electron events the
asymmetry is consistent with zero, for muon events its absolute value in-
creases with momentum and reaches a value around -0.4 above 1 GeV.
5.4. INTERPRETATION OF THE PHENOMENON
It has been suggested that the up-down asymmetry observed for muon
events may result from an excess of downward going muons produced in the
atmosphere and energetic enough to traverse the rock above the detector,
in association with inefficiencies of the outer detector which fails to identify
and reject them.
However, from the pattern of photomultiplier hits it is possible to re-
construct precisely the event vertex and this is required to be at a distance
of at least 2 m from the inner detector walls. No significant excess of events
originating at distances of less than 2 m is observed [28], thus ruling out
this possibility.
Other attempts to explain the up-down muon asymmetry invoke the an-
gular distortions of the cosmic ray flux induced by the Earth magnetic field.
However, these effects are important only in the sub-GeV region and they
are included in the calculations. Furthermore, they have been measured in
the Super-Kamiokande detector itself and found to be in agreement with
predictions [34].
It can be safely concluded, therefore, that Figs. 8 and 9 demonstrate
the existence of a new phenomenon. Its most plausible interpretation is
the occurrence of vI-' oscillations. Since the vl-'/ve ratio at production is
equal to 2, or larger than 2, vI-' - Ve oscillations would induce a large up-
down asymmetry for electrons as well, with more up-going than down-
going electrons, in disagreement with the data. Hence the vI-' predominantly
oscillates to Vr or to a new type of 'sterile' neutrino, which we denote by
V8 •
130 L.DILELLA
Figure 10 shows the region of vlJ - v". oscillation parameters required

to describe the Super-Kamiokande results under the assumption of two-
neutrino mixing. The best fit values are f1m 2 = 3.5 x 10-3 eV2 , sin2 28 = 1,
with X2 = 62.1 for 68 degrees of freedom. The results from the fit are
displayed in Figs. 8 and 9.
.l!I JOO
c (e)
~
•
..
~zoo
11
~ 100
c
0_1 o _"",~-'--{l'-::'.5--'--'-~OL...L..L
1 _I -'
............
O.':'-5..................
Cose Cose
.l!IHiO .l!IHiO
c
~
(e)
I
•
..
QI
'0 100 ~100
11
E 50
::J
C
-{I.5 0 0.5 o -1 -{I.5 0 0.5
Cose Cose
Il: 1.5 , - - - - - - - " I l : 1.5 , - - - - - - - - - - - - ,
(e) Kllllioklulde ~ Super Kllmioklulde
M1lti.QoV
o _l~..LI.::-{I'-:'.5....L...L-'-!Ol...L.L.J....LJO.5'::'-'-.LJ...J o -1 -{I.5 0 0.5
Cose Cose
Figure 8. Zenith angle distributions, as measured in Super-Kamiokande: a) electron-like,

sub-GeV eventsj b) muon-like, sub-GeVj c) electron-like, multi-GeVj d) muon-like,
multi-GeV and partially contained eventsj e) ratio R for multi-GeV events as measured in
Kamiokande [32]j f) ratio R for multi-GeV events, as measured in Super-Kamiokande [28].
The full line represents the expectation in the absence of neutrino oscillation. The dashed
line is the expectation for vI' - V T oscillation with sin 2 (28) = 1, Llm 2 = 3.5 x 10- 3 eV 2 •
e-like
0.&
-0.5
5'
+
~ -~O 'l 10
9 1~-r-r~~~-'~~~",---------,
2. IJ.-like
0.5
FC PC
+' ..•...~' ""., ..
,w•••••••· m•••••m.wmN
'-+- .. ........ ..•..

-0.5
_1~~ __ ~~W-~ __ ~~~ ________ ~
10 . ,
10
Figure 9_ Up-down asymmetry vs energy, as measured in Super-Kamiokande

[28). The hatched bands represent the expectation in the absence of neutrino
oscillation. The dashed line is the expectation for vI-! - V T oscillation with
sin2(20) = 1, Llm 2 = 3.5 x 10- 3 eV 2.
0.8
Figure 10_ Region of vI-! - V T oscillation parameters required to describe the Su-
per-Kamiokande atmospheric neutrino results.
A III' -liB oscillation also gives an acceptable fit with parameters similar
to those required by the III' -liT hypothesis. On the other hand, no oscillation
gives X2 = 175 for 69 degrees of freedom, corresponding to a probability of
132 L. DI LELLA
5 X 10- 9 • The hypothesis of vI' - Ve oscillation is also in disagreement with

the data: the best fit has X2 = 110 for 67 degrees of freedom corresponding
to a probability of 8 x 10- 4 . This hypothesis is also rejected by the results
of the Chooz experiment, as discussed in Section 6.
A definitive proof that the observed anomalies (the low R value and the
distorted muon zenith angle distribution) are the result of neutrino oscilla-
tions can only be obtained by observing the expected oscillatory behaviour
of the v I' signal as a function of LIE over at least a full oscillation cycle (see
Eq. 10). For !:l.m 2 = 3.5 x 10- 3 ey2 signal minima are expected at LIE =
708 (n + 1/2) km/GeY, where n = 0,1,2 ... , thus LIE must be measured
with a resolution significantly smaller than 708 km/GeY. We note that the
distance between two minima (or two maxima) varies as I/!:l.m 2 and the
required LIE resolution also varies in the same way.
In the Superkamiokande experiment the vI' energy in single-ring events
can be measured accurately from the muon range (these events are dom-
inated by quasi-elastic scattering for which the undetected recoil nucleon
has a kinetic energy of typically less than 500 MeY). However, the detector
measures accurately the muon zenith angle, and not that of the incident
neutrino. Thus the neutrino direction is affected by an uncertainty which
is as large as 30° at 1 GeY.
Figure 11 shows the behaviour of the neutrino path length L as a func-
tion of its zenith angle Ov, as given by the equation
where ~ is the Earth radius (6378 km) and h is the altitude at which the
neutrino was produced in the atmosphere (typically of the order of 15 km
with a large uncertainty).
It is clear from Fig. 11 that downward-going neutrinos do not cover a
full oscillation cycle in L / E. Furthermore, for these neutrinos L is poorly
determined because of the large uncertainty on h.
In the Ov region near the horizontal direction (Ov ~ 90°) L varies rapidly
with Ov and it requires a Ov resolution of 1° - 2° to determine L with the
necessary precision. This is possible with neutrinos of very high energies
which, however, are very few and produce muons which are generally not
contained in the detector.
10'
10'
10'
10
-1 -0.8 -0.6 - 0.4 - 0 .2 0 0 .2 0.4 0.6 0.8
cos(zenith angle)
Figure 11. Neutrino path length L vs zenith angle, under the assumption that the
neutrino was produced at an altitude of 15 km.
Finally, for upward-going neutrinos L can be determined with the re-

quired precision using multi-Ge Y muons but a neutrino energy resolution
of the order of 10%, or less, is required in order to reach the resolution in
LIE necessary to observe a full oscillation cycle for tl.m 2 ~ 10- 3 ey2.
A new atmospheric neutrino detector capable of measuring precisely
the incident neutrino direction and energy is needed in order to observe
the oscillatory LIE behaviour for tl.m 2 ~ 10- 3 ey2. The design of such a
detector is not easy. As we shall see in Section 8, the problem is less difficult
for long baseline experiments at accelerators, for which the neutrino path
is fixed and very precisely known.
5.5. UPWARD-GOING MUONS
Charged-current interactions of upward-going 1111 'S in the rock under the

detectors produce upward-going muons which can either stop or traverse the
underground detectors. The study of the zenith angle distribution of these
muons in the interval -1 < cos (}z < 0 has provided additional evidence for
1111 - liz oscillations.
In the Super-Kamiokande detector, upward-going muons are identified
by detecting the entrance and exit point and by requiring that their track
length corresponds to an energy of at least 1.7 GeY [35]. Muons stopping
in the detector are produced by 1111 with < Ell >::;::: 10 GeY, while muons
traversing the detector have < Ell >~ 100 GeY.
134 L.DILELLA
The measured zenith angle distribution of these muons is shown in

Figures 12 a,b. The data disagree with predictions in absence of oscilla-
tions, even if the overall normalization is adjusted to provide the best fit.
However, good agreement is obtained under the assumption of v~ - Vr os-
cillations with full mixing and tl.m 2 values very similar to those required
by the study of muons produced in the detector.
-6
...'&.. a)
UJ
":" 5
UJ
C)I
E 4
...,0
:1
tv)
-u::
'I"'" 3
X
>< 2
:::J
0
-1 -0.8 -0.6 -0.4 -0.2 0
cosS
-
' .1.4 b)
't.fl 1.2
<:'
E
,...u
S2 0.8
6
~ 0.6
-= 0.4
0.2
0
-1 ~.8 -8.6 -8.4 -8.2 0
COSe
Figure 12. Zenith angle distribution of upward through-going muons (a) and up-
ward-going stopping muons (b) as measured in Super-Kamiokande [35], compared with
expectations from no neutrino oscillation (solid line) and II", - I I.. oscillation (dashed line).
Upward-going muons have also been studied in the MACRO experiment

at Gran Sasso [36]. The MACRO detector is a system of horizontal streamer
tube and scintillation counter planes interleaved with absorber plates (slabs
of rock), 9.5 m high and with overall lateral dimensions 77 x 12 m. The
muon direction is determined by time-of-flight. MACRO has very good
detection efficiency for muons near the vertical direction, but the efficiency
decreases rapidly as cosOz approaches zero. The muon energy threshold is
1 GeV and the average parent vI-' energy is 80 GeV. I"V
The zenith angle distribution, as measured in MACRO, is compared

with expectations in Fig. 13. Again, an acceptable fit is obtained only if
vI-' - V T oscillations are assumed, with full mixing and dm 2 close to the
SuperKamiokande value.
....
-. If) ....,............... - .! ........... :.. : ..... , ...... \ ........... .
"" <;I . •
~ ,. ~~
i:: .v ; ~
:...
(J . "............. . ...... .. . .. _ -................._.... ..

.. i -V.!I ·fJ.,~ -0.7 ·0.<1 ·0.' -OA 4.1 ·11.2 - II I (l
1 cos e ...
Figure 13. Zenith angle distribution of upward through-going muons, as measured in
MACRO [36]. The hatched band represents the expectation from no neutrino oscillation
with its uncertainty. The dashed line is the prediction for VI' - v.,. oscillation with full
mixing and ~m2 = 2 x 10- 3 eV 2 •
5.6. vJJ - VT VERSUS V JJ - V 8 OSCILLATIONS
A vI-' - VB oscillation can be distinguished from the oscillation to an 'active'

neutrino by studying neutral-current interactions. Oscillations to an active
136 L.DILELLA
neutrino do not change the rate of these events, whereas VB has neither
charged- nor neutral-current interactions with matter.
In the Super-Kamiokande experiment events consisting of two electron-
like rings are considered to be candidates for the reaction
V +N --t v + 11'0 + N (42)
and the two rings are identified as the photon showers from 11' --t II decay.
Figure 14 shows the two-photon invariant mass distribution [37]. A peak
containing"" 270 events at the nominal11'° mass is clearly visible. The rate
of these events is compared with that of single-ring electron-like events,
which is not affected by neutrino oscillations. This comparison is made by
using the double ratio between the measured and predicted rates in the
absence of oscillation. The result is
e) measured
(11'0 /
(11' 0/)
e pred'lcted = 1.11 ± 0.06 ± 0.02 ± 0.26
where the first error is statistical, the second one represents systematic
uncertainties and the third, dominant error reflects the uncertainty on the
cross-section for neutral-current events resulting in only one detectable 11'0.
A reduction of the order of 30% with respect to unity is expected from
vIJ - VB oscillations.
rnu'S'
c~t
60 for P'Ott
t.Utl\t
se1ccJ1on
Figure 14. Two-photon invariant mass distribution as measured in SuperKamiokande

[37].
Obviously, because of the size of the uncertainty on the 11"0 cross-section

no conclusion can be reached at present (as discussed in Section 9, the
K2K experiment will reduce this uncertainty in the near future). Similarly,
the number of 11"0 events is too small and their correlation to the incident
neutrino direction too weak for a significant analysis of their zenith angle
distribution.
Discrimination between vI' - Vr and vI' - VB oscillations can also be
achieved by exploiting the matter effects which affect neutrinos traversing
the Earth [38]. In the case of vI' - Vr oscillations there are no matter effects
because vI' and Vr have identical interactions with matter. However, in the
case of vI' - VB oscillations the VB does not interact with matter, while the
effective potential for vI' is
where the three terms describe the vI' interaction via Z-boson exchange
with protons, electrons and neutrons, respectively (see the diagram of
Fig. 1a). Because protons and electrons have opposite electric charges one
has Vp = -Ve and
In Eq. (43) N n is the number of neutrons per unit volume, p is the

matter density in g/c:m 3 and Vn is negative for vI' and positive for vI'"
Matter effects are important if IVEI ~ flm2, where E is the neutrino
energy. For flm 2 ~ 3 x 10- 3 eV2 , this implies E ~ 30 GeV. Figure 15
shows a comparison between the measured {}z distribution of muons in
partially contained events (visible energy> 5 GeV, (E) ~ 25 GeV) and
the predictions for vI' - Vr and vI' - Va oscillations [39]. The difference
between vI' - Vr and vI' - VB oscillations expected from matter effects is
clearly visible in the region cos{}z < -004, and the experimental points are
in better agreement with vI' - Vr than vI' - VB oscillations.
A similar conclusion can be drawn from the study of upward through-
going muons [39], as shown in Fig. 16.
138 L.DILELLA
3S
30
!: V'r' tl ! ++r-., .
i:rr~Ij 1":.J
r ~~ _ V tl ... , v"
sr 1 ,\m'=WHe\" (
! sin1 2€l=.1 i
4) !..,.. ,..... '-. ..........;.............................. ....J..., .•.. ~ L ...... , .i.. ....... ~, ........................;
·1 -0.8 -0.6 .0.4 .0.1 0 0.2 0.4 Il.6 0.8 1
cose
Figure 15. Zenith angle distribution of partially contained muon-like events in Su-
per-Kamiokande [39]. The data are compared with ex~ectations from v,. - v~ and v,. - V T
oscillations for full mixing and ~m2 = 3.2 x 10- 3 eV .
'" 2..5 .
0..'1 Sm2 ;: 10'U cV!

&in2 :40;: 1
0:: ' , . . .. .. ' .
·1 ·0.9 ·0.8 ".7 ..e.6 .0.5 -0.4 ..0.3 ·0.2 .0.1
t ro~ ~
Figure 16. Zenith angle distribution of upward through-going muons in Su-

per-Kamiokande [39]. The data are compared with ex~ectations from v,. - v~ and v,. - V T
oscillations for full mixing and ~m2 = 3.2 x 10- 3 eV .
6. Long baseline experiments at nuclear reactors
The Chooz experiment (40] is a fie disappearance experiment sensitive to

oscillations with t1m 2 > 7 x 10- 4 eV2 and sin220 > 0.1. The detector is
installed in an underground site under 115 m of rock at distances of 1114
and 998 m from the two 4.25 GW reactors of the Chooz nuclear power plant
in France.
The detector consists of three concentric vessels. The innermost one

contains 5 tons of Gadolinium-doped liquid scintillator (CH2) which acts
as the fie target. The intermediate vessel, with a mass of 17 tons and also
filled with CH2 liquid scintillator, is used for containment. The outermost
vessel, optically isolated from the two inner ones, contains 90 tons of liquid
scintillator and acts as a veto counter.
The fie are detected by measuring the prompt e+ signal from reaction
(41) followed by the delayed signal from neutron capture. Most neutrons
are captured by Gadolinium, resulting in the emission of ,-rays with a total
energy of '" 8 Me V.
The event rate is 1.1 ± 0.3 d- 1 with both reactors off and 25.5 ± 1.0 d- 1
with both reactors at full power. The ratio between the measured and ex-
pected event rate is 1.010 ± 0.028 (stat.) ± 0.027 (syst.) in the absence of
neutrino oscillation. Figure 17 shows the region of fie - fix oscillation param-
eters excluded by the Chooz experiment at the 90% confidence level. This
region contains the parameter values required to describe the Kamiokande
results in terms of 1/1-' - I/e oscillations [33].
An experiment conceptually similar to the Chooz one has recently started

data taking at the Palo Verde nuclear power plant in Arizona (three reac-
tors with a total power of 10.9 GW). The detector is located at distances
of 750 m from one reactor and of 890 m from the other two in an under-
ground site under 16 m of earth. It consists of 12 tons of Gadolinium-loaded
scintillator surrounded by 1 m thick water shield and by a liquid scintilla-
tor layer to reject cosmic rays. The central detector is segmented into 66
independent cells.
Because of the shallow site the background is much higher in ·the

Palo Verde experiment than in the Chooz one. With all reactors off the
event rate is 32.62 ± 1.02 d- 1 , while it is 39.06 ± 1.00 d- 1 when all reac-
tors run at full power [41]. After 40 days of data taking the experiment
f"V
has not yet reached the sensitivity to oscillations of the Chooz experiment.
140 L.DILELLA
1 r--..-------~~~~------------,
IZl 90% CL Kami okande(multi· GeV)
~ 90% CL Kamiokande(sub+multi·GeV)
10
95% CL
I
10
I
90% CL
10
a2 0.3 a4 as 0.6
Figure 17. Boundary of the ii. - iiz oscillation parameter region excluded by the Chooz
experiment. Also shown is the region of II,.. - II. oscillation parameters allowed by the
results of Ref. [33].
7. Neutrino Oscillation Searches at Accelerators

7.1. IIIJ - lie. iiIJ - iie SEARCHES: LSND AND KARMEN EXPERIMENTS
The Liquid Scintillator Neutrino Detector (LSND) [42] and the KArlsruhe-
Rutherford Medium Energy Neutrino (KARMEN) experiment [43] use neu-
trinos produced in the beam stop of a proton accelerator. LSND has finished
data taking at the Los Alamos Neutron Science Center (LANSCE) at the
end of 1998, while KARMEN is still running at the ISIS neutron spallation
facility of the Rutherford-Appleton Laboratory.
In these experiments neutrinos are produced by the following decay
processes:
(i) 11"+ --+ J.L+vl-' (in flight or at rest);
(ii) J.L+ --+ iil-'e+ve (at rest);
(iii) 11"- --+ J.L-iil-' (in flight);
(iv) J.L- --+ vl-'e-iie (at rest).
The iie yield is very small (of the order of 4 x 10- 4 with respect to iiI-')
because 11"- decaying in flight are a few % of all produced 11"- and only
a small fraction of J.L- stopping in heavy materials decays to vl-'e-iie (11"-
at rest are immediately captures by nuclei; most J.L- stopping in high-Z
materials undergo the capture process J.L-P --+ vl-'n). The neutrino energy
distributions are shown in Fig. 18.
.....
M
"2 lO
::::l
13Q:;1
g
-+
5 W ~ ~ ~ M ~ ~ ~ ~ ~
energy Ev [MeV]
Figure 18. Neutrino spectra from the 7r+ -+ p.+ -+ e+ decay chain at rest.
Table 6 lists the main parameters of the two experiments. For the iiI-' - iie
oscillation search, the iie is detected by reaction (41) which gives a prompt
e+ signal followed by a delayed, signal from neutron capture (the 2.2 Me V
,-ray from the reaction np --+ d, and also, for KARMEN, the 8 MeV line
from ,-rays emitted by neutron capture in Gadolinium, which is contained
in thin layers of Gd2 0 3 placed between adjacent cells).
142 L. DI LELLA
While the LANSeE beam is ejected in rv 500JLs long spills 8.3 ms apart,
the ISIS beam is pulsed with a time structure consisting of two 100 ns long
pulses separated by 320 ns (this sequence has a repetition rate of 50 Hz).
Thus it is possible to separate neutrinos from muon and pion decay from
their different time distributions with respect to the beam pulse.
Table 7 lists preliminary results from LSND [44] and KARMEN [45],
obtained after requiring space and time correlation between the prompt and
delayed signal, as expected from fieP --+ e+n (and for KARMEN requiring
also the time correlation between the e+ signal and the beam pulse). The
LSND result gives evidence for an excess of fie events with a statistical
significance of rv 4.5 standard deviations.
TABLE 6. Parameters of the LSND and KARMEN experiments
LSND KARMEN
Proton beam 800 MeV 800 MeV

kinetic energy
Proton beam current 1000 Jl.A

Single cylindrical 512 cells filled
tankj 1220 PMT'sj with liquid scintillatorj
Detector collection of both cell dim. 18 x 18 x 350 cm
scintillation and
Cerenkov light.
Detector mass 167 tons 56 tons
(mineral oil) (mineral oil)
Event localization timing cell size
Distance from 29 m 17 m
v source
Angle between
v direction and
proton beam
Data taking period 1993-98 Feb. 97-Feb. 99
Protons on target 2.9 X 10 22
TABLE 7. Preliminary results from LSND and KARMEN
LSND KARMEN
e+ energy interval 20-60 MeV 16- 50 MeV
Observed events 70 8
Cosmic ray background 17.7 ± 1.0 1.9 ± 0.1
Total background 30.5 ± 2.7 7.82 ± 0.74
iie signal events 39.5 ± 8.8 < 6.2 (90% C.L)
iiI-' - iie oscillation probability (3.3 ± 0.9 ± 0.5) x 10- 3 < 4.2 X 10- 3 (90% C.L.)
Figure 19 shows the preliminary e+ energy distribution of the 70 events

observed by LSND, together with the distributions expected from back-
grounds and from v/-I - ve oscillations for two different !:lm 2 values. The
region of oscillation parameters describing the LSND result is shown in
Fig. 20, together with the region excluded by previous experiments and by
the present KARMEN result. This figure shows that, if the LSND result is
correct, the only allowed region of 17/-1 - ve oscillation parameters is a narrow
strip with !:lm 2 between 0.2 and 2 eV2 and sin2 20 between 0.002 and 0.04.
§
~. 'o'r'
.R !
~.
·;~~-'-······~·-·;j:6;'··}· -"---!6k-·~·''''······5.6·t.-~-~."·5C:·~-·j
positron energy ( M~ V )
Figure 19. Preliminary e+ energy distribution of the 70 events observed by LSND. Also
shown are the distributions expected from backgrounds (histogram with error bars) and
the expectations from iiI-' - iie oscillations for two different ~m2 values.
144 L. DI LELLA
10
2 1
sin 26
Figure 20. Region v,.. -ve oscillation parameters allowed by the preliminary LSND results
[44]. Also shown are the boundaries of the regions excluded by previous experiments and
by the recent KARMEN results [45].
During the first three years of LSND data taking the target area of the
LANSCE accelerator consisted of a 30 cm long water target located '" 1 m
upstream of the beam stop. This configuration enhanced the probability of
pion decay in flight, allowing LSND to search for vJ.I - Ve oscillations using
vJ.I with energy above 60 MeV. In this case one expects to observe an excess
of events from the reaction
above the expected backgrounds. This reaction has only one signature (a
prompt signal) but the higher energy, the longer track and the directionality
of Cerenkov light help improving electron identification and measuring its
direction.
In this search [46] LSND has observed 40 events to be compared with
12.3 ± 0.9 events from cosmic ray background and 9.6 ± 1.9 events from
machine-related (neutrincrinduced) processes. The excess of events (18.1 ±
6.6 events) corresponds to a vJ.I - Ve oscillation probability of (2.6 ± 1.0) x
10-3 , consistent with the value found from the study of the YeP -+ e+n
reaction below 60 MeV.
7.2. FUTURE SEARCHES FOR vI-' - Ve OSCILLATIONS
The KARMEN experiment will finish data taking in the year 2001. By
then, its sensitivity to vI' - ve oscillations will have improved by a factor of
1. 7 with respect to the present value. However, in case of a negative result
the exclusion region will not fully contain the region allowed by LSND.
A new search for vI' - Ve (or vI' - ve) oscillations is needed, therefore, to
unambiguously confirm or refute the LSND signal.
The Mini-BOONE experiment [47] is a first phase of a new, high sensi-
tivity search for vI' - Ve oscillations (BOONE is the acronym for Booster
Neutrino Experiment). Neutrinos are produced using an 8 GeY, high in-
tensity proton beam from the Fermilab Booster Synchrotron. The beam
consists mainly of vI' from 7r+ decay with a small contamination ('" 0.3%)
of V e , with a broad energy distribution from 0.3 to 2 GeV.
The Mini-BOONE detector will be installed at a distance of 500 m
from the neutrino source. It consists of a 6 m radius spherical tank filled
with mineral oil. Cerenkov light produced in the oil is collected by '" 1500
photomultiplier tubes located on the surface of the sphere. The detector is
surrounded by anticoincidence counters and will use the different pattern
of Cerenkov light expected for muons, electrons and 7r0 to identify these
particles.
For a fiducial mass of 445 tons Mini-BOONE expects to detect", 5 x
105 VJ.I C 12 -+ p,- X events and", 1700 veC12 -+ e- X events in one year.
A vI' - Ve oscillation probability of 0.003 will result in an excess of '" 1500
veC12 -+ e- X events.
IT no oscillation signal is observed, Mini-BOONE will exclude a region
of oscillation parameters which extends to sin2 28 ~ 4 x 10- 4 at large !l.m 2
and to !l.m 2 ~ 0.02 ey2 at full mixing, thus completely excluding the region
presently allowed by LSND. However, if a signal is observed, it should be
possible to measure precisely the oscillation parameters, by using a second
detector at a different distance.
Mini-BOONE will begin taking data in the year 2002.
7.3. vI-' - Vr SEARCHES AT CERN
Two experiments searching for vI' - Vr oscillations have recently completed

data taking at CERN. They both used the wide-band neutrino beam from
the CERN 450 GeV proton synchrotron (SPS). The method adopted by
both experiments consists in detecting T- production with a sensitivity
corresponding to a Vr / v I' ratio much larger than the value expected from
conventional Vr sources in the beam (the main Vr production process is DB
production by the primary protons, followed by the decay DB -+ TVr ). The
observation of T- could only result, therefore, from vI' - Vr oscillations.
146 L. DI LELLA
These experiments are sensitive to !:l.m 2 values above a few ey2. Ac-
cording to the so-called 'see-saw' model [48], the neutrino masses obey one
of the two relations
= m2 2 2
e·. m J.I.. m T'.
m 2u·• m 2e·• m t2
involving either the charged lepton masses (me, mJ.l' m T ) or the Q = 2/3
quark masses (mu, me, mt). Inboth cases one has ml «m2 «m3.
Assuming that the solar neutrino problem is the result of Ve - vJ.I oscil-
lations with small mixing, !:l.m 2 is equal to [m(vJ.I)]2 to a very good approx-
imation, giving m(vJ.I} ~ 3 x 10-3 eY for !:l.m 2 = 10- 5 eV 2. Then, from the
see-saw relations given above the range of values for the V T mass (between
'" 1 and '" 30 eY) is such that the V T could be, at least partially, an im-
portant component of dark matter and vJ.I - VT oscillations can be observed
using vJ.I beams from high-energy proton accelerators and baselines of the
order of 1 km if the mixing angle is not too small.
The two experiments, CHORUS and NOMAD, are installed one behind
the other at a distance of '" 820 m from the proton target. A pair of
pulsed magnetic lenses located after the target produces an almost parallel
wide-band beam of positive hadrons. Neutrinos from 7r or K decay reach
the detectors, while iron and earth shielding absorb surviving hadrons and
range out decay muons. The distance between the proton target and the end
of the decay tunnel is 414 m. Figure 21 shows the expected neutrino energy
spectrum and Table 8 lists its mean energies and relative abundances [49].
The V T 'natural' abundance is estimated to be '" 5 x 10-6 [50].
:" .~
"""
··'i.,
"', v...
............
...........
...... :... D.: -"'1.'"
.......,... -....
".
........ ... ...
~ ~
'"
...
.. '
.~
..
u .. u. u.
NltuilriPl" Erv""rGItVJ
Figure 21. Neutrino fluxes from the CERN wide-band beam.

TABLE 8. Mean energies and relative abundances for v fluxes and

CC interactions at the NOMAD detector [49]. The integrated vI-'
flux is 1.11 X 10- 2 vI-' per proton on target
Flux Charged-current interactions
v type < Ev > Relative < Ev > Relative

GeV abundance GeV abundance
vI-' 23.5 1.00 42.6 1.00

iiI-' 19.2 0.061 41.0 0.0249
Ve 37.1 0.0094 56.7 0.0148
iie 31.3 0.0024 53.6 0.0016
VT '" 35 '" 5 x 10- 6
7.3.1. CHORUS
CHORUS (Cern Hybrid Oscillation Research apparatUS) aims at detecting
the decay of the short-lived T lepton in nuclear emulsion. This technique
provides a space resolution of,..., 1 J.lm, well matched to the average T-
decay length of 1 mm.
The apparatus [51] is shown in Fig. 22. It consists of an emulsion target
with a total mass of 770 kg, followed by an electronic tracking detector
I'V
made of scintillating fibres, an aircore hexagonal magnet, electromagnetic

and hadronic calorimeters and a muon spectrometer consisting of magne-
tized iron toroids interleaved with drift chambers. The hexagonal magnet
provides a field of 0.1 T, over a length of 0.75 m, oriented along the sides of
an hexagone with no radial dependence. It is used to determine the charge
and momentum oflow-energy particles with a resolution cr(p)jp 20% for I'V
momenta between 2 and 10 GeV.

Neutrino events with a J.l- or a negatively charged hadron are selected
and the tracks are followed back to the exit point from the emulsion tar-
get. The method is illustrated in Fig. 23. It relies on special emulsion sheets
mounted between the target and the fibre tracker (these sheets are replaced
every few weeks during the run). With the reconstruction accuracy of the
fibre tracker the track position on the special sheet is predicted within an
area of 360 J.lm x 360 J.lm. In this area, because of the short exposure time,
one finds, an average, 5 muon tracks which are rejected by angular mea-
surement. The search is then continued in an area of 20 J.lm x 20 J.lm into
the emulsion target, with negligible background despite the long exposure
time of the target (2 years).
148 L. OI LELLA
CHORUS
.,001 baI:
i DUOIl 8p«tramel:er
Figure 22. Layout of the CHORUS detector.
Vlewaf
microsc~p.
Figure 23. Expected configuration of a typical V T charged-current interaction in the

CHORUS emulsion. In this example the r- decays to Jj-vTv,...
Events in which the neutrino interaction point is found in the emul-

sion target are analysed to search for T- decay, which is identified by the
presence of a change in direction (kink) of a negatively charged track, as ex-
pected from one-prong decays. No other charged leptons must be observed
at the primary vertex and the transverse momentum (pr) of the selected
particle with respect to the T candidate direction is required to be larger
than 0.25 GeV (to eliminate strange particle decays). Negative tracks with
impact parameters larger than rv 8 J.Lm are also considered as possible

secondary particles from r - decay.
Negative muons with momentum p < 30 GeV or negative hadrons with
1 < p < 20 GeV are considered as possible r- decay products. IT observed,
the kink is required to be within 3.95 mm from the interaction point for
muon tracks, and 2.37 mm for hadron tracks. The emulsion scanning pro-
cedure to locate the interaction point and to reject events incompatible
with r- decays is fully automatic. The remaining events undergo a com-
puter assisted eye scan to confirm the presence of a r decay. This step is
necessary because the development process deforms the emulsion and the
tracks become distorted. It must be verified, therefore, that the decay kink
or the large impact parameter are genuine and are not the result of local
distortions. This is achieved by scanning and measuring background tracks
close to the event under consideration (in the CHORUS experiment these
tracks are mostly straight-through muons from SPS beams).
CHORUS took data between May 1994 and the end of 1997. Table 9
summarizes the present status of the analysis [52]. No r- candidate has
been observed yet.
The dominant background in the one-muon channel is the production
of charmed particles from iJlJ interactions
iJlJ N -+ J.L+ D- X
followed by the decay D- -+ J.L- + neutral particles. These events are
dangerous only if the J.L+ is not identified. Their contribution to the data
sample analysed so far is 0.24 ± 0.05 events.
In the muonless channel the background from charm production amounts
to 0.075 ± 0.015 events. A more serious background is the interaction of neg-
ative hadrons with nuclei producing only one outgoing negatively charged
particle with no evidence for nuclear break-up (these interactions are called
'white kinks'). The rate of such events is affected by a large uncertainty and
is estimated to be 0.66 ± 0.66 events.
TABLE 9. Status of the CHORUS analysis [52]
One-muon events Muonless events

Expected number of events 458.6 x 10 3 116 x 10 3
Fraction scanned so far 54% 47%

Events with identified v interaction 126.2 x 10 3 19.4 X 103
N,. for < PIJT > =1 4876 1137

150 L.DILELLA
For a particular" decay channel, i, the expected number of events is
! ~v"
given by
N: = BRi (E)PI-'TuT A~ €~ dE
where: BRi is the decay branching ratio; ¢v" (E) is the vI-' energy spectrum;
PI-'T is the oscillation probability; U T is the cross-section for ,,- production;
A~ is the acceptance and reconstruction efficiency (including the vertex-
finding efficiency); and €~ is the efficiency of the decay search.
Similarly, the expected number of one-muon events from vI-' charged-
current interactions is given by
At large tl.m 2 values PI-'T is constant and one can use average values:
Ni = N P BRi < A~ > < €i > (44)

T I-' I-'T < UI-' > < AI-' > T
The ratio / < ul-' > has the value 0.53. For the" -t J.L decay
channel < A~ > / < AI' > is 1.075 because of the requirement PI' <
30 GeV which suppresses vI-' charged-current interactions more than" -t J.L
decays. The efficiency of the decay search is estimated by simulations and
its value, €~ ~ 0.37, is verified experimentally using J.L- J.L+ events from
vl-'N -t J.L- n+ X followed by n+ -t J.L+ decay.
Muonless ,,- decays include the ,,- -t h- and ,,- -t e- channels, where
h- is a charged hadron, and also ,,- -t J.L- events in which the J.L- was not
identified. To allow an easy combination of the results from the one-muon
and muonless channels, an 'equivalent number of muonic events' is defined
for the muonless sample using the relation
Neq = N 2:i < A~ >< €~ > BRi

I-' I-' < A~ >< €~ > BRIJ '
where the sum extends to all muonless " decay channels and NI-' is the
corresponding number of one-muon events from vI-' charged-current inter-
actions. With this trick the oscillation probability PI-'T can be determined
for muonless events using Eq. (44) with i = J.L and replacing N I' by N~q.
The expected values of N"., as obtained by setting PIJT = 1 in Eq. (44)
are given in the last row of Table 9. The 90% confidence upper limit on PI-'T
is
PIJT < 2.38/NT = 4.0 X 10- 4
where 2.38 is the Poisson upper limit of a null result at the 90% confidence
level taking into account a systematic uncertainty of 17% on NT [53]. The
region of oscillation parameters excluded by this result under the assump-

tion of two-neutrino mixing is shown in Fig. 24 together with the results
from previous experiments [54].
~
E531 i
8
<l :, i!1(:
J ,
I..'
I !
~O{lllAlP
10 2 (Jq:r II) ' I
l ,
10
NOMAD+CHORUS
v ->v
I.L t
90% C.L.
CDHS
.1
10 L4
~~~~L3~~~~L.'2~~~~.'1~~~~
10 10 10 10 1
sin2 28
Figure 24. The ~m2 - sin2(20) plane for v,. - V T oscillation.The regions excluded by
CHORUS and NOMAD, and by their combined results [55], are shown together with
the results of previous experiments [54]. Full lines: CHORUS (for two different statistical
methods); dashed lines: NOMAD and combined result.
A more stringent limit, PI!T < 2.6 X 10- 4 , would be obtained [55] using
the method recently proposed by Feldman and Cousins [56].
CHORUS is expected to reach the upper limit PI!T < 10- 4 is no event
is found after the completion of the analysis.
7.3.2. NOMAD
NOMAD (Neutrino Oscillation MAgnetic Detector) is designed to search
for vI! - VT oscillations by observing T- production using kinematical criteria
[57], which require a precise measurement of secondary particle momenta.
The main detector components are [58] (see Fig. 25):
- drift chambers (DC) used to reconstruct charged particle tracks and
also acting as the neutrino target (fiducial mass 2.7 tons, average f"V
density 0.1 gjcm3 , radiation length 5 m); f"V

152 L.DILELLA
- nine independent transition radiation detectors (TRD) for electron

identification, interleaved with additional drift chambers;
- an electromagnetic calorimeter (ECAL) located behind a 'preshower'
detector (PRS);
- a hadronic calorimeter (HCAL);
- large-area muon chambers.
Figure 25. Side view of the NOMAD detector.
DC, TRD and ECAL are located inside a uniform magnetic field of
0.4 T perpendicular to the beam direction.
The NOMAD experiment aims at detecting r- production by observing
both leptonic and hadronic decay modes of the r-. The decay r- -+ vre-ve
is particularly attractive because the main background results from Ve
charged-current interactions which are only'" 1.5% of the total number
of neutrino interactions in the target fiducial volume and have an energy
spectrum quite different from that expected from vI-' - Vr oscillations (see
Fig. 21 and Table 8). The selection of this decay relies on the presence of
an isolated electron in the final state and on the correlation among the
lepton transverse momentum (Pf), the total transverse momentum of the
hadronic system (P¥) and the missing transverse momentum (P;p) (only
the momentum components perpendicular to the beam direction can be
used because the incident neutrino energy is unknown). In the case of Ve
charged-current events Pf is generally opposite to p¥
and IPP I is small (it
should be exactly zero if the momenta of all secondary particles were mea-
sured precisely and the target nucleon were at rest). On the contrary, in
r- -+ vre-ve decays there is a sizeable IPPI associated with the two outgo-
ing neutrinos. Furthermore, in a large fraction of events PP is at opposite
azimuthal angles to ff¥, in contrast with Ve charged-current interactions for
which large values of I~I result mostly from hadrons escaping detection
(in these cases the azimuthal separation between p!j. and ~ is small) .
The selection of hadronic r- decays relies on the observation of a
hadron, or of a collimated system of hadrons, consistent with r- decay
and well isolated from the other hadrons in events containing no primary,
isolated lepton. A powerful variable to reject neutral-current events is the
transverse momentum of the candidate hadron(s) from r- decay with re-
spect to the total visible momentum:
~ .... )2
(p T)2 _ (Ph' P
h p2'
where Ph is the momentum of the candidate hadron(s) from r- decay and

P is the total visible momentum. For neutral-current events P is the total
momentum of the hadronic jet and QT is the transverse momentum of a
particle (or system of particles) in the jet. This is generally much smaller
than the value expected for hadron(s) from r- decay which are well iso-
lated from the hadronicjet produced in V T charged-current interactions (see
Fig. 26).
;~~Jo 0.5 l l.5 2 2.5 3 3.5 4 4.5 5

QT GeV
Figure 26. Expected NOMAD distribution of the variable QT (see text) for simulated
neutral-current (Ne), charged-current (ee) interactions and r- -+ IIT 7f'- decays.
II,..
A particularly dangerous background in the search for hadronic r- de-

cays results from vf.1 or Ve charged-current interactions in which the pri-
mary, high-PT lepton was not identified. This background is reduced to a
tolerable level by rejecting all the events containing a high-PT negative par-
ticle which could not reach the sub detectors providing lepton identification
(TRD, PRS, ECAL, muon chambers).
In practice, for most r- decay channels the method to separate the
signal from backgrounds uses ratios of likelihood functions. These functions
are approximated by products of probability density functions of kinematic
variables. These are obtained from large samples of simulated events, after
corrections to take into account differences between simulated and real data,
as described in Ref. [59].
154 L. DI LELLA
For the r- -+ e- decay channel the analysis uses two likelihood ratios,
Ael and Ae2, which are used to separate the signal from neutral-current and
Ve charged-current background, respectively. Figure 27 shows the event dis-
tribution in the InAel, In Ae2 plane for backgrounds and signal. The framed
region in the upper-right corner (the 'signal box') contains a sizeable frac-
tion of r- -+ e- events and little background. The predicted background
amounts to 5 .3~g:: events [60].
8 r-------- . - . - - - - - , - - - -- - - -, - , -- - - - ,
/Iw .;
~ " ';
S~:·
o '':
-2
~~" ': ': .
ISOLATION
' - - - ->
o 5 o 5
Log. IIkelih ood ratio (tiNe)
Figure £7. Scatter plot of In Ae2 vs In Ael for (a) simulated v,.. neutral-current interac-
tions; (b) simulated Ve charged-current interactions; (c) simula.ted r - -+ e-v.,.iie events;
(d) NOMAD data.
The analysis is not allowed to look at the data in the signal box until
a robust background prediction has been provided. When the signal box is
opened, 5 events are found (see Fig. 27d), in agreement with the predicted
background.
Table 10 shows a summary of backgrounds and efficiencies for separate
analyses of deep-inelastic scattering events (DIS) and low-multiplicity (LM)
events reported in Ref. [60] (DIS events, as are defined by the requirement
pH > 1.5 GeV). For all channels there is good agreement between the
observed number of events and the background prediction. The resulting
90% confidence level upper limit using the method of Ref. [56] is [60]
PIJT < 4.2 X 10- 4 ,
which corresponds to the exclusion region shown in Fig. 24.

The CHORUS and NOMAD limits can be combined using the method
of Ref. [56]. The combined limit is [55]
PIJT < 1.3 X 10-4 •
The corresponding exclusion region is outlined in Fig. 24.
TABLE 10. Summary of NOMAD results [60]. N,. is the

number of expected r- events for PlAT = 1
Analysis NT Estimated Observed

background number of
events
r- -+ e- DIS 4110 5.3::g:~
r- -+ h-(mrO) DIS 2232 6.8 ± 2.1 6
r- -+ p- DIS 2547 7.3::~:~ 8
r- -+ 11"-11"-11"+ DIS 1180 6.5 ± 1.1 5
r- -+ e- LM 859 5.4 ± 0.9 6
r- -+ h- LM 357 6.7 ± 3.0 5
r- -+ p-LM 457 5.2 ± 2.4 7
r- -+ 11"-11"-11"+ LM 108 O.4::g:~ 0
8. Long baseline experiments at accelerators

Long baseline experiments at accelerators extend the sensitivity of searches
for vlJ - Vz oscillations to dm 2 as low as 10- 3 eV2 using vlJ beams of well
known properties which can be monitored and varied if needed. The main
goal of these experiments is to verify that the atmospheric neutrino results
discussed in Section 5 are indeed associated with oscillations, to establish
the nature of the oscillation and to measure its parameters.
Table 11 shows a list of parameters for the three existing projects.
156 L.DILELLA
TABLE 11. Long baseline projects
Project Accelerator Location of Distance < Ell > Status

far detector (Km) GeV
K2K KEK 12 GeV Kamioka 250 1.4 Start
proton synchrotron mine April 1999
NuMI Fermilab 120 Ge V Soudan 730 16 Start
Main Injector (MI) mine or lower 2002
NGS CERN 450 GeV Gran Sasso 732 30 not yet
SPS Lab or lower approved
8.1. K2K
The K2K project [61] uses neutrinos from the decay of 1r and K mesons
produced by the KEK 12 GeV proton synchrotron and aimed at the Su-
perKamiokande detector at a distance of 250 km. The beam consists mainly
of 1/1-" with vI-' and 1/e contaminations of 4% and 1%, respectively. The 1/1-'
energy spectrum is shown in Fig. 28.
For a run of 1020 protons on target (3 years, corresponding to an ef-
fective data taking time of 12 months) the total number of events in the
SuperKamiokande fiducial volume (22.5 Ktons of water) is expected to be
3451/1-' charged-current, 120 neutral current and 41/e charged-current inter-
actions in the absence of neutrino oscillations. The beam energy is below
threshold for r- production (Ell ~ 3.5 GeY), so no search for r- appear-
ance is possible.
Figure 29a shows the expected distortion of the vI-' flux at 250 km for a
oscillation with dm 2 = 3.5 x 10-3 ey2 and full mixing. Such a distortion
vI-'
can be detected by comparing the energy distribution of beam-associated
muon-like events in SuperKamiokande with the distribution measured in a
similar, 1 Kton water detector located at 300 m from the neutrino source on
the KEK site (see Fig. 30). No oscillation effects are expected at this short
distance and the number of events is much larger (I'V 4 X 105 1/1-' charged-
current interactions in a fiducial volume of 21 tons). Figure 31 shows the
region of oscillation parameters which is excluded at the 90% confidence
level if no significant difference is observed between the near and far detec-
tor.
.'
~.('...... r!!~
~:,
•..
:,:~j?$:;;
r:
'J
'~!'c-J
"
::>
.;:;"
.. .
. . ... ,
~
Figure 28. Expected VI-' energy spectrum at the SuperKamiokande detector in the
absence of neutrino oscillation.
O.B
0 .6
c
.2
0.4
0.2
/ a)
......
o
X
:l
[L
0.8
0.6
0.4
0.2 b)
05 1. 5 2 2 .5 3 3 .5 4.5
Neutrino energy (GeV)
Figure 29. Expected flux modulation from vI-' disappearance in the K2K (a) and NUMI
project (b), for a two-neutrino oscillation with .6.m2 = 3.5 x 10- 3 eV 2 and full mixing.
The energy values where minima and maxima occur are proportional to .6.m2 •
158 L.DILELLA
Figure 30. Layout of the near detector of the K2K project.
90%C.L.
10
.,
~
..!.
(II
E
<l ·2
10
-.3
10
Figure 31. Region of V,. - VT or v,. - v. oscillation paramaters excluded at the 90%
confidence level if no oscillation signal is detected by the K2K experiment after three
years of data taking.
The near detector includes a system of scintillating fibres in water, a

lead glass calorimeter and a muon range telescope to monitor and mea-
sure precisely the vI" vI' and Ve energy spectra and space distributions.
The near detector will also measure the cross-section for 11"0 production in
neutral-current interactions. As discussed in Section 5.6 the knowledge of
this quantity helps understanding if the apparent disappearance of atmo-
spheric vl"s is the result of oscillations to an active or to a sterile neutrino.
In the K2K project the expected number of single-11"° events in Super-
Kamiokande is only 25 for a three-year run. This number is too small to
provide any significant information on the issue of active versus sterile neu-
trinos from the K2K experiment itself.
K2K has started data taking in April 1999 and has detected the first in-
teraction of a beam neutrino in the fiducial volume of the Super-Kamiokande
detector.
8.2. NUMI AND THE MINOS EXPERIMENT
The NuMI project uses neutrinos from the decay of 11" and K mesons pro-
duced by the new Fermilab Main Injector (MI), a 120 GeY proton syn-
chrotron capable of accelerating 5 x 1013 protons with a cycle time of 1.9 s.
The expected number of protons on target is 3.6 x 102o /y. The decay pipe
is 675 m long.
The neutrino beam will be aimed at the Soudan mine in Minnesota (an
inactive iron mine) at a distance of 730 km from the proton target. The
beam will consist primarly of vI" with 0.6% Ve' Figure 32 shows the expected
energy distributions of vI' charged-current events for three different neutrino
beams which correspond to different tunes and locations of the focusing
elements. For the high energy beam the total number of events is '" 3000/y
for a detector mass of 1000 tons.
The expected distortion of the vI' flux for an oscillation with 8m 2 =
3.5 x 10- 3 ey2 and full mixing is shown in Fig. 29b. It is clear that such a
distortion can be best detected using the lowest energy beam of Fig. 32.
The MINOS experiment [62] will use two detector, one (the 'near de-
tector') located at Fermilab, the other (the 'far detector') located in a
new underground hall to be built at the Soudan site at a depth of 713 m
(2090 m of water equivalent). Both detectors are iron-scintillator sandwich
calorimeters with a toroidal magnetic field in the iron plates.
The far detector (Fig. 33) has a total mass of 5400 tons and a fiducial
mass of 3300 tons. It consists of magnetized octagonal iron plates, 2.54 cm
thick, interleaved with active planes of 4 cm wide, 8 m long scintillator
strips providing both calorimetric and tracking information.
160 L.DILELLA
> 400 ,-------------------,

~
"- 50 m target pile + 675 m d."oy pipe
'-350 I'·.
II>. ,.... r· ~ '. ~'.;'"
"-
c 300 "L; Perfect Focusing
~
"-
~200
c
~ :"',
~200
()
...
',"'1_.
'..
o0 • 10 10 20 25 30 ~ 40
[(II.) GeV
Figure 32. Neutrino interaction energy spectra predicted for different focusing conditions
at the Soudan location of the NuMI project. 'Perfect focusing' is the ideal case of all
secondary 11'+ and K+ being focused into a pencil beam.
MINOS (Man Injoctor NeuOi no O scillation S....ch )

Far Detector
25.800 Dl ' AcljyeDetedor PI", ...
WLS fi bu readout
Magnetized Fe Plates
4S6 Layers x 2.54 cm Fe
5.4 kT Total Mass
Figure 33. Sketch of the MINOS detector.
The near detector has a total mass of 920 tons an a fiducial mass of 100
tons. It will be installed 250 m dowstream from the end of the decay pipe.
The comparison of vI-' charged-current event rate and energy distribution
in the two detectors will be sensitive to oscillations which can be detected
with a statistical significance of at least four standard deviations over the

full parameter space currently suggested by the atmospheric neutrino re-
sults (see Fig. 34). The oscillation parameters will be measured precisely
over most of this region.
[C Muon Neutrlno Ols~ppe~r~nce Test
10 kton-yeClrs
2.0! Nu Fjux~u:;:.onc.:....:T.:.....-~..d
'">
~ 10"2
'"E
<J
Figure 34. Excluded region (90% confidence) and 40" discovery region for a 10 kton x
y exposure of the MINOS experiment from a comparison of the 1/,.. CC event spectra in
the far and near detector.
The measurement of the ratio between neutral- and charged-current

event rate (NC/CC) is very important because it will be used to discrimi-
nate between vJj - liT and lIJj - lis oscillations. In the former case NCICC
is larger in the far detector, while for lIJ.I - lis oscillations it has the same
value in the near and far detector.
The MINOS experiment should begin data taking at the end of the year
2002.
8.3. THE CNGS PROJECT
The CNGS project (CERN Neutrinos to Gran Sasso) has not yet been
approved. It consists in aiming a neutrino beam from the CERN 450 Ge V
SPS to the Gran Sasso National Laboratory in Italy at a distance of 732
km. The three existing underground halls at Gran Sasso, under '" 4000 m
of water equivalent, are already oriented towards CERN and ICARUS [63],
a 600 ton detector suitable for oscillation searches, will start operation in
the year 2001 to search for proton decay and to study atmospheric and
solar neutrinos.
162 L.DILELLA
If approved before the end of 1999, the CNGS beam will be operational
in the year 2005. It will be used for V T appearance experiments, for which
a detector in a 'near' location should not be necessary. The rate of VT
charged-current interactions from VII - VT oscillations is given by
(45)
where A is a normalization constant proportional to the detector mass,

¢v,. (E) is the VII energy spectrum at the detector, PIIT(E) is the VII - VT
oscillation probability and U T (E) is the cross-section for V T charged-current
interactions. The integration lower limit is set by the energy threshold for
r- production, 3.5 GeV.
Under the assumption of two-neutrino mixing, PIIT(E) is given by
Eq. (10). For a large fraction of the tl.m 2 interval suggested by the at-
mospheric neutrino results the condition 1.27 tl.m 2 L / E < 1 holds for L =
732 km and E > 3.5 GeV. In this case Eq. (45) can be approximated as
.2 (20) (tl.m)
NT = 1.61 A sm 222/
L ¢v,. (E)uT(E) dE
E2 . (46)
In this approximation NT varies as (tl.m 2)2. In addition, because at large

distances ¢v,. (E) decreases as L -2, NT does not depend on L. However, the
background from conventional neutrino events is proportional to L -2, so
the ratio between the r- signal and the background varies as L2.
In the approximation of Eq. (46), the neutrino beam must be designed
with the goal of maximizing the integral of Eq. (46) which does not depend
on tl.m 2. A preliminary beam design, based on a 1000 m long decay tunnel,
is described in Ref. [64]. The design has been improved [65] to optimize
the r- production rate. The VII mean energy is 17 GeV and the rate of
VII charged-current interactions is '"V 2448/y for a detector with a mass of
1000 tons (this value corresponds to 4.5 X 1019 protons on target, which is
a realistic figure for a one-year run after the shut-down of LEP).
The rates of other neutrino events relative to VII events are 0.007, 0.02
and 0.0007 for Ve, fill and fie, respectively. Table 12 lists the expected yearly
rates of r- events [65] for three different tl.m 2 values.
TABLE 12. Yearly

rate of r- events vs
dm 2 for a 1000 ton
detector [66J (1 year
= 4.5 x 10 1 protons
on target).
dm 2 (eV2) Nr
10- 3 2.48
3 X 10- 3 21.7
5 x 10- 3 58.5
ICARUS [63] is a new detector concept based on a liquid Argon Time

Projection Chamber (TPC) which allows three-dimensional reconstruction
of events with spatial resolution of the order of 1 mm.
Primary ionisation electrons drift in very pure liquid Argon over dis-
tances of the order of 1.5 m and are collected by wire planes which provide
two of the three coordinates and measure the ionisation, while the third
coordinate along the drift direction is determined by measuring the drift
time.
Liquid Argon has a density of 1.4 g/cm3 , a radiation length of 14 cm and
an interaction length of 84 cm. ICARUS is also an excellent calorimeter,
with an expected resolution (jE/ E = 3%/VE and 13%/VE for electromag-
netic and hadronic showers, respectively (E in GeY).
A 600 ton ICARUS module is presently being constructed. The cryostat
cold volume (534 m3 ) is 19.6 m long and 4.2 m high. Three additional
modules will be built if the operation of the first module is successful.
ICARUS will search for T appearance using kinematical criteria similar
to those used in the NOMAD experiment (see Section 7.3.2). However, for
the T- -+ e- decay channel a background rejection power in excess of 104
was needed in NOMAD, while in ICARUS a rejection power of rv 102 is
sufficient because of the much smaller number of events. This allows looser
selection criteria and the detection efficiency for T- -+ e- events becomes
f'V50%. With four modules one expects rv 10 T- events/y for a 1/1-' - 1/7"
oscillation with 11m 2 = 5 x 10- 3 ey2 and full mixing, to be compared with
a background of 0.25 events.
The T identification from other T decay channels is presently under
study. Since the T- production rate at low !l.m 2 is proportional to (!l.m 2 )2,
a detector consisting of four ICARUS modules should be sensitive to oscil-
lations with 11m 2 ~ 2 x 10- 3 ey2 after a running time of four years.
164 L. DILELLA
Very recently, a new detector (ICANOE) with a total active mass of

9,300 tons has been proposed [66]. This detector consists of several ICARUS
modules interleaved with conventional calorimeters made of magnetized
iron.
Another interesting detector concept for a V T appearance search is
OPERA [67]. The detection of one-prong T decays is performed by mea-
suring the T- decay kink in space, as determined by two track segments
measured with very high precision in nuclear emulsion.
Figure 35 illustrates the OPERA concept. The main component of the
target are 1 mm thick Pb plates where most neutrinos interact. One-prong
T decays occurring in the 3 mm gap between the emulsion detectors ES1,
ES2 are expected to result in observable kinks. ES1 and ES2 consist of two
50 p,m thick emulsion layers glued to a 100 p,m plastic foil. The 3 mm gap
between ES1 and ES2 is filled with a very low density spacer to which ES1
and ES2 are glued to ensure that their relative positions are stable and
precisely known.
- \"I.
J mm 3 IT""
1
so 00 50 <1-1111)
Fi!Jtlre 35. Schematic structure of the OPERA target.
The OPERA detector is arranged in 'bricks', each consisting of rv 30 Ph-

ES1-ES2 layers with transverse dimensions 15 x 15cm2 , for a total mass of
rv 800 tons. Magnetized iron toroids and conventional trackers will be used
to identify p, - and to minimize the background from the decay of charm

particles produced by vJJ charged-current interactions. Conventional track-
ers are located after each plane of bricks to identify the brick where the
neutrino interaction took place. This brick is immediately removed from
the detector for scanning and measurement of the emulsion layers.
The global T- detection efficiency is estimated to vary between 0.29 at

low ~m2 and 0.33 at ~m2 ;:::: 10- 2 eV 2 . For a run of one year (4.5 x10 19
protons on target) the background is expected to be 0.4 events, mostly from
charm production and decay in events in which the primary 11- - was not
identified. The number of expected T - events for full mixing is listed in
Table 13 for different ~m2 values.
TABLE 13. Number of detected

r- in OPERA per 4.5 x 10 19
protons on target (1 year) (from
Ref. [67])
t..m 2 (eV2) Detected N- /y
l.0 X 10- 3 0.5
3 X 10- 3 4.5
5.0 X 10- 3 12.0
The OPERA concept is being also considered at Fermilab in the frame

of the NuMI project. For the highest energy version of the NuMI beam (see
Fig. 32) the yearly T- production rate in a 1000 ton detector is a factor of
'" 2 higher than in the NGS project, because of the much higher number
of protons on target (3.6 x 1020 for NuMI, 4.5 x 1019 for NGS). However,
the high energy neutrino beam is not the best choice for MINOS if ~m2 is
below", 5 x 10- 3 eV 2 . It is not clear, therefore, if the two experiment will
be compatible.
9. Summary and Conclusions
As discussed in the previous sections, studies of solar and atmospheric

neutrinos and oscillation searches at accelerators have provided evidence or
hints for neutrino oscillations. If these results are interpreted in terms of
two-neutrino mixing, the following regions of oscillation parameters in the
(sin 220, ~m2) plane are suggested:
- several 'islands' with ~m2 in the range 4 x 10- 11 - 5 X 10- 10 eV2

and large mixing angles (the 'vacuum oscillation' solutions to the solar
neutrino problem);
- two islands with ~m2 ;:::: 10- 7 eV2 or 4 x 10- 5 eV2 and large mixing
angles, or one with ~m2 ;:::: 6 x 10- 6 eV2 and small mixing angle (the
MSW solutions to the solar neutrino problem);
166 L.DILELLA
- an island with flm 2 in the range 10-3 - 10-2 ey2 and large mixing
angle, dominated by vI' - vr or vI' - VB oscillation (the solution to the
atmospheric neutrino problem);
- a narrow strip with flm 2 between 0.2 and 2 ey2 and sin220 between
0.002 and 0.04, required to describe the iiI' - iie oscillation signal
claimed by LSND.
From this information one can draw the following conclusions:
- four neutrino states are needed to describe simultaneously the solar
and atmospheric neutrino measurements and the LSND result, because
with three neutrinos there are only two independent flm 2 values. IT
the need for a fourth neutrino is confirmed, then this neutrino must
be sterile (no coupling to Wand Z bosons);
- unless neutrinos are degenerate in mass, flm 2 is equal to the square
of the mass of the heavier neutrino to a very good approximation. In
this case the heaviest neutrino has a mass of at most 1.4 eY. Hence
neutrinos are not the main component of dark matter in the Universe.
By the middle of the next decade, more data from experiments presently
running and results from experiments just beginning or in preparation will
provide answers to several crucial questions. In particular, we expect to
learn if the solar neutrino problem is due to oscillations and to know the
oscillation parameters with much smaller uncertainties than the present
ones.
We expect also to know rather precisely the oscillation parameters re-
sponsible for the atmospheric neutrino problem, either from further data
on atmospheric neutrinos, or from long baseline experiments at accelerators
(or from both). We should also know if the dominant oscillation is vI' - Vr
or vI' - VB from various indirect measurements or from the observation of
Vr appearance.
Finally, we shall definitely know if the iiI' - ve oscillation signal observed
by LSND is real.
In the longer term, if neutrino oscillations are confirmed, the elements
of the mixing matrix need to be measured. This task requires neutrino
beams at least two orders of magnitude more intense than present ones.
The most promising idea to achieve this goal is based on high-energy muon
storage rings with long straight sections pointing to a neutrino detector. The
advantage of such 'neutrino factories' is that they provide beams of precisely
known composition (50% vI" 50% ve from 1'- decay, or 50% vI" 50% Ve from
1'+ decay), and also precisely calculable fluxes and energy spectra [67]. With
such beams, backgrounds are very low for some appearance searches, such
as vI' - Vr oscillations detected by observing the T- --t e- decay channel. In
addition it may become possible to measure CP violation in the neutrino
sector.
The construction of neutrino factories represents a challenge for accel-

erator technology.
Acknowledgements
I wish to express my sincere thanks to Jean-Jacques Aubert, Director of the

1999 Cargese Summer School on Particle Physics, for inviting me to give
these lectures and for organizing a very successful and stimulating meeting.
I am grateful to Dominique Escalier for her help. Finally, last but not least,
I thank Genevieve Prost for typing and editing these lectures.
References
1. Particle Data Group, Eur. Phys. J. C3 (1998) 122.
2. For a review see D.O. Caldwell, Neutrino Dark Matter, hep-ph/9902219.
3. V.M. Lobashev, Direct Search for the Neutrino Mass, in Proc. of the XVIII Physics
in Collision Conference, Frascati (Italy), June 17-19, 1998.
4. K Assamagan et aI., Phys. Rev. D53 (1996) 6065.
5. R. Barate et aI., Eur. Phys. J. C2 (1998) 395.
6. L. Baudis et aI., Limits on the Maiorana neutrino mass in the 0.1 eV range, hep-
ex/9902014.
7. B. Pontecorvo, Sov. Phys. JETP 26 (1968) 984.
8. Z. Maki, M. Nakagawa and S. Sakata, Prog. Theor. Phys. 28 (1962) 870.
9. L. Wolfenstein, Phys. Rev. DI1 (1978) 2369.
10. S.P. Mikheyev and A.Yu. Smirnov, Nuovo Cim 9C (1986) 17.
11. J.N. Bahcall, Rev. Mod. Phys. 50 (1978) 881; Rev. Mod. Phys. 59 (1987) 505;
J.N. Bahcall and R.K Ulrich, Rev. Mod. Phys. 60 (1988) 297;
J.N. Bahcall and M. Pinsonneault, Rev. Mod. Phys. 61 (1995) 781.
12. J.N. Bahcall, S. Basu and M. Pinsonneault, Phys. Lett. B433 (1) 1998;
J .N. Bahcall, P.I. Krastev and A.Yu. Smirnov, Phys. Rev. D58 (1998) 6016.
13. R. Davis, Prog. Part. Nucl. Phys. 32 (1994) 13.
14. B.T. Cleveland et aI., Astrophys. J. 496 (1998) 505.
15. S. Turck-Chieze and I. Lopez, Astrophys. J. 408 (1993) 347.
16. W. Hampel et aI., Phys. Lett. B441 (1999) 127.
17. V.N. Gavrin NucL Phys. B (Proc. SuppL) 11 (1999) 20.
18. Y. Fukuda et aI., Phys. Rev. Lett. 81 (1998) 1158; Phys. Rev. Lett. 82 (1999) 1810;
M.B. Smy, Solar Neutrino Energy Spectrum with SuperKamiokande, presented at
DPF99, Los Angeles, January 5-9, 1999.
19. J.N. Bahcall, M. Pinsonneault, S. Basu and J. Christensen-Dalsgaard, Phys. Rev.
Lett. 18 (1997) 171.
20. N. Hata and P. Langacker, Phys. Rev. D49 (1994) 3622.
21. Y. Suzuki, Solar Neutrinos, presented at the XIX Int. Symp. on Lepton and Photon
Interactions at High Energies. (LP-99), Stanford, August 9-14, 1999.
22. A.B. Mac Donald, Nucl. Phys. B (Proc. Suppl.) 77 (1999) 43.
23. L. Oberauer, Nucl. Phys. B. (Proc. Suppl.) 77 (1999) 48.
24. A. Suzuki, Nucl. Phys. B. (Proc. Suppl.) 77 (1999) 171.
25. G. Barr, T.K Gaisser and T. Stanev, Phys. Rev. D39 (1989) 3532;
V. Agrawal et aL, Phys. Rev. D53 (1996) 1314;
M. Honda et aI., Phys. Lett. B248 (1990) 193; Phys. Rev. D52 (1995) 4985.
26. KS. Hirata et aI., Phys. Lett. B205 (1988) 416; Phys. Lett. B280 (1992) 146.
27. D. Casper et aI., Phys. Rev. Lett. 66 (1991) 2561;
R. Becker-Szendy et aI., Phys. Rev. D46 (1992) 3720.
168 L.DILELLA
28. Y. Fukuda et aI., Phys. Rev. Lett. 81 (1998) 1562;

Phys. Lett. B433 (1998) 9; Phys. Lett. B436 (1998) 33;
M. Messier, Atmospheric Neutrinos at SuperKamiokande, presented at DPF99, Los
Angeles, January 5-9, 1999.
29. Ch. Berger et aI., Phys. Lett. B227 (1989) 489; Phys. Lett. B245 (1990) 305;
K. Daum et aI., Z. Phys. C66 (1995) 417.
30. M. Aglietta et aI., Europhys. Lett. 8 (1989) 611.
31. W.W.M. Allison et aI., Phys. Lett. B391 (1997) 491;
E.A. Peterson, Nucl. Phys. B (Proc. Suppl.) 77 (1999) 111.
32. S. Kasuga et aI., Phys. Lett. B374 (1996) 238.
33. Y. Fukuda et aI., Phys. Lett. B335 (1994) 237.
34. T. Futagami et aI., Phys. Rev. Lett. 82 (1999) 5194.
35. Y. Fukuda et aI., Phys. Rev. Lett. 82 (1999) 2644;
Y Totsuka, Evidence for Neutrino Oscillation Observed in Super-Kamiokande, pre-
sented at the Nobel Symposium, Stockholm, 20-25 August 1998.
36. M. Spinetti, Atmospheric Neutrino Results from MACRO, presented at the VIII
Int. Workshop on Neutrino Telescopes, Venice, Italy, February 23-26, 1999.
37. K. Scholberg, Atmospheric neutrinos in Super-Kamiokande, presented at the VIII
Int. Workshop on Neutrino Telescopes, Venice, Italy, February 23-26, 1999.
38. P. Lipari and M. Lusignoli, Phys. Rev. D58 (1998) 073005.
39. A. Mann, Atmospheric Neutrinos, presented at the XIX Int. Symp. on Lepton and
Photon Interactions at High Energies (LP-99), Stanford, August 9-14, 1999.
40. M. Apollonio et al., hep-ex/9907037 (to be published in Phys. Lett. B).
41. Y. Wang, Recent Results from Palo Verde and Future Prospects, presented at the
XXXIV Rencontres de Moriond on Electroweak Interactions and Unified Theories,
Les Arcs, France, 13-20 March 1999.
42. C. Athanassopoulos et al., Nucl. Intrum. Methods A388 (1997) 149.
43. B. Zeitnitz, Prog. Part. Nucl. Phys. 32 (1994) 351.
44. G. Mills, Results from LSND, presented at the XXXIV Rencontres de Moriond on
Electroweak Interactions and Unified Theories, Les Arcs, France, 13-20 March 1999.
45. T. Jannakos, Latest Results from the Search for V" ~ v. Oscillations with KAR-
MEN2, presented at the XXXIV Rencontres de Moriond on Electroweak Interactions
and Unified Theories, Les Arcs, France, 13-20 March 1999.
46. C. Athanassopoulos et al., Phys. Rev. C58 (1998) 2489.
47. E. Church et al., A proposal for an experiment to measure vI' -v. oscillations and VI'
disappearance at the Fermilab Booster: BOONE, 7 December 1997 (unpublished).
48. T. Yanagida, Prog. Theor. Phys. B135 (1978) 66;
M. Gell-Mann, P. Ramond and R. Slansky, in Supergravity, eds. P. van Nieuwen-
huizen and D. Freedman (North Holland, Amsterdam, 1979) 315.
49. G. Collazuol et aI., Neutrino Beams: Production Models and Experimental Data,
CERN-OPEN-98--{)32.
50. M.C. Gonzales-Garcia and J.J. Gomez-Cadenas, Phys. Rev. D55 (1997) 1297;
B. Van de Vijver and P. Zucchelli, Nucl. Instr. and Methods A385 (1997) 91.
51. E. Eskut etal., Nucl. Instr. Methods A401 (1997) 7.
52. The CHORUS Collaboration, hep-ex/9907015.
53. R.D. Cousins and V.L. Highland, Nucl. Instr. Methods A320 (1992) 331.
54. E531: N. Ushida et aI., Phys. Rev. Lett. 57 (1986) 2897;
CHARM-II: M. Gruwe et aI., Phys. Lett. B309 (1993) 463;
CCFR: K.S McFarland et aI., Phys. Rev. Lett. 75 (1995) 3993;
CDHS: F. Dydak et aI., Phys. Lett. B134 (1984) 281.
55. R. Petti, private communication.
56. G.J. Feldman and R.D. Cousins, Phys. Rev. D57 (1998) 3873.
57. C. Albright and R. Shrock, Phys. Lett. B84 (1979) 123.
58. J. Altegoer et aI., Nucl. Instr. Methods A404 (1998) 96.
59. J. Altegoer et al., Phys. Lett. B431 (1998) 219;
P. Astier et al., Phys. Lett. B453 (1999) 169.

60. D. Gibin (NOMAD Collaboration), Search for II", - liT oscillations in the NOMAD
experiment, presented at the Int. Europhys. Conf. on High Energy Physics (HEP99),
Tampere, Finland, 15-21 July 1999.
61. K. Nishikawa, Nucl. Phys. B (Proc. Suppl.) 77 (1999) 198.
62. S. Wojcicki, Nucl. Phys. B (Proc. Suppl.) 77 (1999) 182.
63. A. Bettini et al., Nucl. Instr. Methods A332 (1993) 395;
J.P. Revol et al., A search program for explicit neutrino oscillations at long and
medium baselines with ICARUS detectors, ICARUS-TM-97/0l (5 March 1997).
64. G. Acquistapace et al., The CERN neutrino Beam to Gran Sasso, CERN 98-02
(19 May 1998).
65. A. Ereditato et al., Towards the optimization of the CNGS neutrino beam for II", -
liT appearance experiments, ICARUS-TM/98-13, OPERA 980722-01 (August 10,
1998);
J.L. Baldy et at., The CERN Neutrino beam to Gran Sasso (NGS), CERN-SL/
99-034 (May 1999).
66. F. Arneodo et al., ICANOE: Imaging and CAlorimetric Neutrino Oscillation Ex-
periment, INFN/AE-99-17; CERN/SPSC 99-25 (1999).
67. K. Kodama et al., The OPERA liT appearance experiment in the CERN-Gran Sasso
neutrino beam, CERN SPSC 98-25 (October 9, 1999).
68. S. Geer, Phys. Rev. D51 (1998) 6989.
PHYSICS OF CP VIOLATION AND RARE DECAYS
T.NAKADA
CERN EP-Division
CH-1211 Geneva 23, Switzerland
and
Institute of High Energy Physics
University of Lausanne
CH-l015 Dorigny, Switzerland
On leave from PSI, CH-5232 Villigen-PSI, Switzerland
1. Introduction
Symmetries are one of the most fundamental concepts for understanding

the laws of nature leading to conserving quantities. Unexpected violations
of symmetries indicate some dynamical mechanism beyond the current un-
derstanding of physics.
Parity violation was discovered in 1957 [1] in nuclear (3 decays and pion
and muon decays [2]. In the charged current interaction of the standard
electroweak theory, parity and charge conjugation symmetries are maxi-
mally violated due to the V - A structure [3]. All the experimental results
up to now are in full agreement with the theory.
A surprising discovery of the CP violating KL --+ 7T+7l' - decays [4] was
made in 1964. The neutral kaon system still remains to be the only place
CP violation has been seen. The Standard Model with three Fermion fam-
ilies can accommodate all the observed CP violation phenomena through
the complex quark mixing matrix, Cabibbo-Kobayashi-Maskawa (CKM)
matrix [5]. However, no real precision test has been made due to the large
uncertainties in evaluating the effect of hadronic interactions.
Interest in CP violation is not limited to elementary particle physics.
It is one of the three necessary ingredients to generate observed excess
of matter over antimatter in the universe [6]. The amount of CP violation
which can be generated by the Standard Model appears to be insufficient for
171
1.-1. Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments, 171-202.
172 T . NAKADA
explaining the observed matter-antimatter asymmetry in the universe [7],

giving a strong motivation to search for new physics.
For CP violation in some B meson decay channels, the Standard Model
can make precise predictions with little influence from the strong interac-
tions. Those channels can be used to test the predictions quantitatively
to look for a sign of new physics. Also in the B meson system, CP viola-
tion is expected in many decay modes. The pattern of CP violation allows
us to make a systematic qualitative comparison with the Standard Model
predictions. Therefore, it is now widely accepted that the B-meson system
provides in future an ideal place for testing the Standard Model for CP
violation [9].
In this article, we first derive the formalism [10] describing the particle
antiparticle system, with and without CP violation. Three different mecha-
nisms which can generate CP violation are clearly classified, together with
experimental observables which identify contributions from the different
mechanisms. Then, CP violation in the neutral kaon system is analysed in
this formalism. After a brief discussion on the Standard Model description
for CP violation in the neutral kaon system, we proceed to the neutral
B meson system. Following the discussion on some Standard Model pre-
dictions, some thoughts are made how the situation could change if there
exists new physics contributing in the B meson system.
2. Description of Particle Antiparticle System

2.1. BASIC FORMALISM
Let IPO) and JPO) be the states of a neutral pseudoscalar particle po_
meson and its antiparticle pO-meson at rest, respectively. They have definite
flavour quantum numbers with opposite signs: F = +1 for pO and F = -1
for pO. Both states are eigenstates of the strong and electromagnetic inter-
action Hamiltonian, i.e.
where mo and mo are the rest masses of pO and pO, respectively. The pO and
pO states are related through CP transformations. For stationary states, the
T transformation does not alter them, with the exception of an arbitrary
phase. While CP is a unitary operation, T is an anti unitary operation.
In summary, we obtain
CP IPO) = ei Bcp JPO) and CP JPO) = e- i Bcp IP O)

(1)
T IPO) = eiBTIPO) and T IPO) = eiilTIPO)
PHYSICS OF CP VIOLATION AND RARE DECAYS 173
where the O's are arbitrary phases, and by assuming CPT IP O) = TCP IP O)
it follows that
20cp = eT - OT .
Since T is antiunitary, it follows that
Tc = c*T
where c is any complex number. If we define
Tla) = la), TI(3) = I~)

antiunitary operation has to give
On the other hand,
hence
We can then conclude
i.e. when the T operator changes the direction of the operation, it must be
complex conjugated.
If strong and electromagnetic interactions are invariant under the CPT
transformation, which is assumed throughout this paper, it follows that
mo = mo·
Now we switch on an interaction, V , and the P can decay into final states
f with different flavours (I6.FI = 1 process) and po and po can oscillate
to each other (I6.FI = 2 process) . Thus, a general state 17/J(t)) which is a
solution of the Schrodinger equation
i :t l7/J(t)) = (Hst + Hem + V) 17/J(t)) (2)
can be written as
17/J(t)) = a(t)IPo) + b(t)iPo) + LCf(t)lf)

f
where the sum is taken over all the possible final states f and a(t), b(t)
and Cf(t) are time dependent functions; la(t)12, Ib(t)12 and ICf(t)1 2 give the
174 T. NAKADA
fractions of pO, pO and f at time t respectively. Since the weak interaction

is much weaker than strong and electromagnetic interactions, perturbation
theory can be applied in order to solve equation 2. Also with the help of the
Wigner- Weisskopf method, which neglects the weak interactions between
the final states [8], and we obtain
z. ata( a(t)
b(t)
) _ (a(t)) _ (
- A b(t)
.
- M - z'2
1') ( a(t)
b(t)
) (3)
where the 2 x 2 matrices M and l' are often referred to as the mass and
decay matrices.
The elements of the mass matrix are given as
(4)
where P stands for the principal part and the index i = 1(2) denotes pO (pO) .
Note that the sum is taken over all possible intermediate states common to
pO and pO for i =1= j.
The elements of the decay matrix are given by
rij = 27r L(iIVlf) (fIVIJ)6(mo - Ef) (5)

f
The sum is taken over only real final states common to pO and pO for i =1= j.
If the Hamiltonians are not Hermitian, transition probabilities are not
conserved in decays or oscillations, i.e. the number of initial states is not
identical to the number of final states. This is also referred to as the break
down of unitarity. We assume from now on that all the Hamiltonians are
Hermitian, i.e.
la(t)12 + Ib(t)12 + L Icrl 2 = 1,
f
and also
Clearly la(t)12 + Ib(t)j2 decreases as a function of time, hence A is not

Hermitian.
Since the CP operator changes a particle state into an antiparticle state,
the following relation can be obtained if V is invariant under the CP trans-
formation, i.e. (CP)-l V CP = V:
Since the T operator induces complex conjugation, which is equivalent to

interchanging a bra-state and a ket-state, the following relation can be
obtained if V is invariant under the T transformation:
By combining the two, we obtain for the CPT invariant case:
CPT: All = A22 .

For a rigorous proof, equations 1, 4 and 5 are used.
It follows that
-if All #- A22 , i.e. Mn #- M22 or rn #- r 22 :
CPT and CP are violated
(6)
-if lAd #- IA211, i.e. sin(tpr - tpM) #- 0:
T and CP are violated .
where tpM = arg (M12 ) and tpr = arg (r12 ). Note that CP violation in
the mass and decay matrices cannot be separated from CPT violation or
T violation.
While there is no fundamental reason to respect CP and T symmetries,
it can be shown based on only few basic assumptions that no self consistent
quantum field theory can be constructed that does not conserve CPT sym-
metry [11]. Therefore, we restrict our further discussion to the case where
CPT symmetry is conserved: i.e.
Mll = M22 == M , rn = r22 == r

thus
An = A22 == A .
Differential equation 3 can be reduced to
d2 a(t) da(t)
-;Ji2 + 2i A ~ + A12A21 - A
( 2) a(t) = 0 (7)
for a(t), and a general solution is given by
where C± are arbitrary constants which can only be defined by the initial
condition. For b( t) , we obtain
176 T.NAKADA
which can be used once a(t) becomes known.

Insertion of a(t) into equation 7 leads to
from which the eigen-frequencies are obtained as
by solving where
(8)
and
(9)
For an initially pure pO state, we have a(t) = 1 and b(t) = 0 at t = 0,
i.e. C+ = C_ = 1/2, and the solution is given by
a(t)IPO) + b(t)IPo)
f+(t)IPo) + (f-(t) iPO) (10)
where
and ( is
(12)
The two states IP +) and IP _) are the eigenstates of A± and are given by
(13)
For an initially pure pO state, we have
IPO(t)) = ~ f-(t) lPo) + f+(t) iPo) (14)
= J\~1(12 (1P+)e-i,\+t-IP_)e- iLt ) (15)

While p± have definite masses and decay widths (as seen from equations
11 and 15), pO and pO do not and they oscillate to each other (see equations
10 and 14) .
2.2 . CP CONSERVING CASE
If V remains invariant under the CP transformation, from equations 1, 4

and 5 it follows that
M 12 -- M 21 e - i28cp -
-
M*12 e-i28cp
thus
arg M12 = -Bcp + n7r,
and
r 12 -- r 21 e-i28cp -
-
r*12 e-i28cp
thus
where nand n' are arbitrary integer numbers.

For (, we have
where nil is an arbitrary integer number. The two mass eigenstates IP +)

and IP _) become CP eigenstates
The mass and decay width eigenvalues, equations 8 and 9, become
and
By examining various combinations of n, n' and nil, we can show that

the following four possibilities exist:
1. n=even, n' =even: C P = +1 state is heavier and decays faster ,
2. n=even, n'=odd: CP = +1 state is heavier and decays slower ,
3. n=odd, n'=even: CP = +1 state is lighter and decays faster ,
4. n=odd, n' =odd: C P = + 1 state is lighter and decays slower.
178 T.NAKADA
I) 1m 2) 1m
Re Re
3) 1m 4) 1m
Re - 6cp Rc
Figure 1. Relative phase relations for M12, Ft2, and CP transformation phase Bcp when
CP is conserved: 1) CF = +1 state is heavier and decays faster, 2) CF = +1 state is
heavier and decays slower, 3) CF = +1 state is lighter and decays faster, 4) CF = +1
state is lighter and decays slower.
Figure 1 illustrates the phase relations in a pictorial way. The choice of nil
does not alter the conclusion and nil = 0 can be adopted without any loss
of generality. In this case, IP +) is the CP = +1 state.
2.3. CP VIOLATING CASE
Let us consider the time dependent decay rate for the initial pO decaying
into a CP eigenstate f, given by 1(flVlp°(t)W, and that for the initial pO
decaying into f, given by I(f IV iPo (t) W:
R,(t) ex 11+(t) I' + (~: '11-(t)I' + 2R [(~: 1~(t)1-(t)l (16)
Rr(t) ex Ar
Ar 2If+(t)1 2+"(
1112 If-(t)1 2+ RP~
2 [*( A; *
A;f+(t)f-(t) 1(17)
where the instantaneous decay amplitudes are denoted by Ar == (fIVIPO)
etc. and equations 10 and 14 are used.
Since Rf(t) and Rf(t) describe the CP conjugated processes to each

other, any difference between the two is an clear proof of CP violation. As
seen from the first terms of equations 16 and 17, CP violation is generated
if IAfl =F IAfl· This is called CP violation in the decay amplitudes .
l.From the second terms of Rr(t) and Rf(t), it can be seen that CP
violation is generated if I( 1=F 1 even if there is no CP violation in the decay
amplitudes. From equations 11 and 15, it is clear that the oscillation rate
for pO --+ pO is different from that for pO --+ pO if 1(I =F 1, thus this is called
CP violation in the oscillation.
The third term can be expanded into
2R ((~:) R [J~(t)f-(t)J - 2~ ((~:) ~ [J~(t)f-(t)J

for Rf(t) and
1~2R ((~:) R [J~(t)f-(t)J + 1(~2 ~ ((~;) ~ [J~(t)f-(t)J

for Rf (t). If CP violation in pO _po oscillation is absent, the first terms are
identical. Even in that case, if
~ ((~;) =F 0
CP violation is still present. Since the process involves the decays of pO
(PO) from the initial pO (PO) and decays of the pO (pO) oscillated from the
initial pO (pO) into a common final state, it is referred as CP violation
due to the interplay between the decays and oscillations.
If CP violation in po_po oscillation is small, i.e. (1(1 - 1)2 < < 1, we can
derive
1 sin(cpr - CPM)I « 1
from equation 12, where CPr = arg r 12 and cP M = arg M12 as already de-
fined. By introducing Ib.r /MI « 1 as
CPr - CPM = n7r - b. r /M (18)
where n is an integer number, the following two approximations are possi-
ble:
a) CPr = argr12 base
( ~ {l- 2IM12I!r121b.r/M [(_l)n+1 +i 2IMd J} e-i'Pr

41Md 2 + Ir d 2 Ird
m± M±(-l)nIMd
r± r±lrd
180 T.NAKADA
b) <PM = arg M12 base
( {I + 2IM121Ir12l~r/M [(_l)n + i Ir12 1]} e-i'PM (19)

+ Ir12l 2
~
41M1212 21 M121
m± = M±IMI21
r± = r ± (-ltlrI21 .
3. Neutral Kaon System

3.1. ADAPTATION OF FORMALISM
Now we adapt the above developed formalism to the neutral kaon system.
As described later, observed CP violation in the KO-Ko oscillation is very
small. The two mass eigenstates are called Kg and KL with corresponding
masses and decay widths referred to as mg, mL, rg and rL respectively
and they are known to be mg < mL and rg > rL· Therefore, M12 and Ft2
is almost antiparallel to each other, thus n = 1 in equation 18.
Since the kaon decay properties are experimentally well measured, enough
information is available to calculate r 12 from the data, as described in Sec-
tion 3.5. We therefore adopt the <P r base given in the previous section.
It follows that
(20)
where the small parameter f is given by
f=
and
IKg) (21)
(22)
l.From the measured lifetimes [12],
Ts =rg1 = (0.8934 ± 0.0008) x lO-10 s
and
TL =rL1 = (5.17 ± 0.04) x lO-8 s
i.e.
tlF = Fs - FL = (1.1174 ± 0.0010) x lO lD s-l
and the mass difference,
tlm == mL - ms = (0.5301 ± 0.0014) x 1010 ns- 1

we obtain,
IMdlFd = 0.24966 ± 0.00004
41Md 2 + IFd 2
and
211M1,I = 0.9488 ± 0.0026 .
F12
Since the lifetime of KL is much longer than that of Ks, it is possible

to produce a KL beam. Therefore, many kaon experiments have been done
using KL beams.
3.2. CP VIOLATION IN OSCILLATIONS
The CPLEAR experiment observed CP violation in the KO-Ko oscillation

by measuring the difference in the oscillation rates between KO ---+ KO and
KO ---+ KO. The initial neutral kaons were produced by pp annihilations:
pp ---+ KOK-7r+ and ---+ KOK+7r -, where the initial flavour can be defined by
the charge sign of the accompanying kaon. Semileptonic decays were used
in order to determine the flavour at the moment of the decay. Since the
KO contains an s-quark (and KO an s-quark), KO (KO) can decay only into
e+7r- v (e-7r+P) instantaneously. Therefore, the initial KO (KO) can produce
the final state e-7r+P (e+7r-v) only through the KO ---+ KO (KO ---+ KO)
oscillation. From the two measured time dependent decay rates, Re- (t)
and Re+ (t), an asymmetry
is constructed as shown in Figure 2. Using equations 10, 14 and 20 , it follows

that
1 _1(1 4
AT(t) =
1+ (
I 14 = 4RE
and from the measured AT(t) = (6.6 ± 1.6) x 10- 3 [13],
1(1 = 0.9967 ± 0.0008 -=J 1
is obtained exhibiting a clear sign of CP violation and T violation in the

KO -KO oscillation.
182 T.NAKADA
.,.
t
; ... 0.04
«
0.03
0.02
0.01
0
-0.01
-0.02
2 4 6 8 10 12 14 16 18 20
Neutral-kaon decay time [Ts]
Figure 2. Measured rate asymmetry between the initial RO decaying into e+ll'-v and
the initial K O decaying into e-ll'+i/ as a function of the decay time in units of TS by the
CPLEAR experiment. The solid line is obtained by fitting a constant value.
The parameter I( 1can also be measured from the semileptonic branching

fractions of KL by the lepton sign asymmetry: using equations 22 and 20,
we obtain [12]
B(KL --t e+1I'-v) - B(KL --t e-1I'+v)

B(KL --t e+1I'-v) + B(KL --t e-1I'+v)
1 _1(1 2
= 1 + 1(12 = 2~€
= (3.27 ± 0.12) x 10- 3
where e can be e or J1. and B stands for a branching fraction.

Using all the measurements, we obtain
~€ = (1.64 ± 0.06) x 10- 3
and
arg € = (43.50 ± 0.08t.
3.3. CP VIOLATION DUE TO DECAYS AND OSCILLATIONS
Since the two-pion final state is a CP eigenstate with CP = +1, KL decay-

ing into 11'+ 11'- is a CP violating decay. This was indeed the first observed
sign of CP violation. A commonly used CP violation parameter 17+- is
defined as
_ (1[+1[- IVIKd 1- (1+-+- (23)
f/+- = (1[+1[- IVIKs) 1 + (A+_
II+-
where equations 21, 22 are used and A+_ and A+_ denote the KO and
KO ---+ 1[+1[- decay amplitudes respectively.
The parameter f/+- can be measured from the time dependent decay
rates for the initial KO and KO into 1[+1[-. From equations 11 and 15, the
two rates are given by
and
where <p+- is the phase of f/+- and t is the Ks-KL average decay width.
The second term is CP violating KL decays and the third term is due to the
interference between the Ks decay and CP violating KL decay amplitudes.
Figure 3 shows [14] the measured R+_(t) and R+_(t) together with the CP
asymmetry defined as
where the interference term is well isolated. At around t = lOTS, the Ks

decay rate is reduced to the level of the CP violating KL decay rate, thus
the asymmetry becomes very large.
This direct comparison between the two CP conjugated processes illus-
trates another straightforward demonstration of CP violation in the neutral
kaon system. From the asymmetry, the value of f/+- is measured to be [14J
If/+-I = (2.264 ± 0.035) x 10- 3 , <p+- = (43.19 ± 0.60t

which leads to
8' ((~:=) = -(3.099 ± 0.048) x 10- 3
exhibiting that CP violation due to the interference between the decay and
oscillation is present.
184 T.NAKADA
0.1
-0.1
-0.2
-0.3
-0.4
8 10 12 1. 16 18 20
-0.5 ~2............4~~6""""-:!8~1"'=0"""""12':'-'-'1:'-:4"""""'1'="6.......",,18~20·
Neutral-kaan decoy time ['T,l Neutral-kaon decay time ('T,l
Figure 3. The time dependent rate distributions for the initial RO (solid circles) and KO
(open circles) decaying into 1('+71"- as a function of the decay time in units of TS obtained
by the CPLEAR experiment. The rate asymmetry is also shown.
3.4. CP VIOLATION IN DECAYS
The two-pion final state can be in a total isospin state of I = 0 or I = 2.

The I = 1 state is not allowed due to Bose statistics. Using the isospin
decomposition, we can derive the KO and KO decay amplitudes to 11"+ 11"- to
be
and
Using CPT symmetry and the S-matrix, the KO and KO decay amplitudes
can be related and it follows that
where ao and a2 are the KO decay amplitudes into 211"(1 = 0) and 211"(1 = 2)
states due to the short-range weak interactions and 80 and 82 are the 11"-11"
scattering phase shifts for the I = 0 and I = 2 two-pion configuration at
...;s = mK respectively. It is important to note that the two-pion scattering
is totally dominated by the elastic scattering at the energy scale of the kon
mass. Similarly for the nOn o final state, we have
ADD -I'faoe + j'fa e

i 00 2 i 02
ADD -I'faoe (oO+fJCP-BT) + j'fa'2ei

i (ih+fJcp-BT )
As seen from the amplitudes , B(Ks --7 nOn O)/ B(Ks --7 n+n - ) would be
0.5 if a2 = O. Since the measured ratio is rv 0.46 [14], we can conclude that
Iad aD I < < 1. It follows that
A+_ = (1 _ 2(') e- i (2<Po+BT- OCP) (24)
A+_
where the parameter (' is given by
(' = _1-la2Isin(<P2 _ <p0)e i (7r/2+tl 2 -oO ) (25)

v'2 aD
and <po, 2 = arg aD, 2·

As seen from equation 24, CP violation in the decay amplitude, IA+-II-
IA+-I, is present if ~(' I- o. From equation 25, this is possible only if
sin(<p2 - <po) I- 0 and sin(62 - 60) I- 0 .
i.e. both the weak and strong phases have to be different for the I = 0
and I = 2 decay amplitudes. More generally, there must be two processes
leading to the identical final state and both the strong and the weak phases
must be different between the two processes in order to generate CP vio-
lation in the decay amplitudes. It should be noted that from the measured
n-n scattering phase shift values, we have [15]
arg(' = (43 ± 6t
Using equations 20 and 24, it follows that
where the approximation is made assuming that the phase difference be-
tween F12 and AoAo is small, which will be justified later. From equation
23, 'f/+- can be derived to be
186 T.NAKADA
Similarly the CP violation parameter for the 7r 0 7r 0 decay channel, 1]00, is

given by
Thus, we expect CP violation parameters to be different between the 7r+ 7r-

and 7r 0 7r 0 decay modes if E' i= O. It has been shown by four recent experi-
ments, NA31 [16], E731 [17], KTeV [18] and NA48 [19],
Irt+_12 = 1.0127 ± 0.0028

1]00
i.e. CP violation in the decay amplitude is present in the neutral kaon
system. If we neglect (cpr + 2cpo + OT - Ocp), it follows that
3.5. PHASE OF DECAY MATRIX
As seen from equation 5, evaluation of F12 involves the decay final states
which are common to KO and KO, which are 27r(I = 0), 27r(I = 2), 37r(I =
1), 37r(I = 2) and 37r(I = 3) states:
Ft2 ~ L A;7r(I)A27r(I) + L A;7r(I) A 37r(I).

1=0,2 1=1,2,3
The contribution from the decay amplitude to the 27r(I = 2) state
is suppressed by the 1:11 = 1/2 rule and the small measured value of E'.
The contribution from the three-pion decay amplitudes are suppressed by
FL/ Fs and the measured upper limits for the CP violation parameter for
the 7r+ 7r-7r0 and 7r0 7r0 7r 0 final states. In conclusion, the phase of F12 is
essentially given by the phase of the A o amplitude, and it can be expressed
as
so that
ICPr+ 2cpo + -OT - Ocpl < 0(10- 5 ) .
Thus ICPr + 2cpo + OT - Ocpl « lEI, justifying the approximations made
before.
3.6. THE STANDARD MODEL DESCRIPTION
In the framework of the Standard Model [20], the short range contribution
to KO-Ko oscillation is obtained from the box diagrams (Figure 4) to be
Mf~x = - G~f~~;7r~Km~ [1]la;S(XC) +2rt2acatE(xc,Xt) +1]3a;S(xd]

s
d
...
w -
t
t c u
c u
01(
~w
,
d
s
0s
d
.....
n
C
-----------------0
W
W
__________________
..... 1
nl
CI -
d
K K K K
Figure 4. The box diagrams contributing to the KO-Ko oscillations.
where GF is the Fermi constant, iK, BK and mK are the decay constant, B
parameter and mass for the K-meson respectively and mw is the mass of
the W-boson. The QCD correction factors are denoted by "71 = 1.38 ± 0.20,
"72 = 0.57 ± 0.01 and "73 = 0.47 ± 0.04 and Sand E are known functions of
the mass ratios, Xi = mflm'f.v for top (i=t) and charm (i=c). Note that
S(Xe) ~ 2.4 X 10- 4 , S(Xt) ~ 2.6, E(xe, xc) ~ 2.2 X 10- 3 (26)
for me = 1.25 GeV Ie, mt = 274 GeV Ie and mw = 80 GeV Ie [12J. The
parameters o"c and O"t are the combination of the elements of the Cabibbo-
Kobayashi-Maskawa quark mixing matrix (CKM-matrix),
0"c= Vcs Vcd * and O"t = vts vtd *. We adopt the following approximation of
the CKM matrix using the parameters introduced by Wolfenstein [21J:
(27)
where where p = p(l - >.2/2) and it = "7(1 - >.2/2). The parameter >.
is known from the light hadron decays to be 0.221 ± 0.002. From the B-
meson decays, lVebl = 0.0402 ± 0.0019 and lVub/Vcbl = 0.090 ± 0.025 are
measured [12], giving A = 0.823 ± 0.042 and J p2 + "7 2 = 0.41 ± 0.11.
The B-parameter takes in account the difference between (OIHwK±) and
UIHwIKO) where (01 is the hadronic vacuum state and (II is the common
quark states between KO and KO. The theoretical evaluations for this value
vary between 0.5 and 1.
In addition to Mf2ox , there are large contributions from long range in-
teractions Mh
R , which are difficult to evaluate. Therefore, theoretical pred-
ication for M12 = Mf:fx + Mr2R cannot be given. The long range interaction
188 T.NAKADA
K K
Figure 5. Gluonic and electromagnetic penguins contributing to the 1(0 -+ 27r decays.
involves only the light flavours and its contribution to M12 is real in the
CKM phase convention; the imaginary part of M12 is generated only by the
box diagram. Therefore we can derive
. ( ) _ ~M12 _ 2~Mf:r
sm rpM - IM12I - t:lm .
In the CKM phase convention, F12 can be approximated as real. Therefore,

it follows that
~Mbox
~E = _ 12
2t:lm
Although there are considerable uncertainties to evaluate numerically this
expression, the currently allowed range of the Wolfenstein parameters, A,
A, p and 'fJ gives a consistent value of ~E with the experimentally measured
value.
Prediction of E' requires an accurate evaluation of the phase difference
between ao and a2. For the ao amplitudes, the tree, the gluonic penguin
and the electroweak penguin diagrams contribute. Only the tree and elec-
troweak penguin diagrams make contributions to the a2 decay amplitude.
All the penguin diagrams are shown in Figure 5. Not only the short range
interactions, but also the hadronic matrix elements with long range inter-
actions have to be evaluated in the calculations. This makes the numerical
determination of E' very difficult. Within the theoretical uncertainties, val-
ues of E' calculated with the currently allowed range of A, A, p and 'fJ are
consistent with the data.
3.7. CP VIOLATION IN RARE DECAYS
Experimental detection of KL -t 7r°v/J is clearly very challenging. The final

state is a CP eigenstate with C P = +1. Therefore, observation of this decay
is a sign of CP violation. In the Standard Model, the decay is generated by
penguin diagrams or box diagrams as shown in Figure 6.
Since the final state consists with only one hadron, long range strong
interactions do not play a role and the decay amplitUdes can be denoted
as
(7fOvvIHw IKO)
(7fovvIHwI K o)
Unlike for the KO -+ 27f decays, 1>7rOVV = arg a 7r oVV could be very different
from 1>0, so that we could have a situation
Isin (1)r + 2 1>7rOVV + (}cP - OT) I Isin (2 1>7r OVV - 21>0)1

» lEI·
The KL decay amplitude then becomes
(7f ovvIHwI K L) a:;; [1 - (1 _ 2E)e- i (24)rr ovv- 2 !f>ol]

~ V2 i la7rovvl sin(2 1>7rOVV - 21>0) .
Using isospin symmetry, the hadronic matrix element of the KO -+ 7fOVV

decay amplitude and that of the K+ -+ 7f+e+v decay amplitudes can be
related as
(7fOIHwIKO) = (7foIHwI K +) .
This allow us to express the branching fractions for KL -+ 7fOVV using the
branching fractions for K+ -+ 7foe+v as [20]
fL
°
B(K + -+ 7f e + v ) -TL ---'=--....:,-~.:..::..:..-,-~~
30'.2 [8'(V't;Vid)X(mt)]2
T+ lVus 1227f2 sin4 8 w
B(K+ -+ 7f Oe+v) TL 3 0'.2 [~(mt)]2 A4 ,\8(1 _ ,\2/2)2r?
T + 27f2 sm4 8w
~ 3 X 10- 11
~~... ----~
.... ~
t cu
---.~~ d
~
K K
d
"
Figure 6. The box and penguin diagrams generating RO -+ 7r°VV decays.
190 T.NAKADA
where X is a known function and 8w is the weak mixing angle. Since the
hadronic matrix element is taken from the data, the theoretical uncertain-
ties in this determination is very small. Also the imaginary part of the
amplitude is dominated by the short range interactions which can be reli-
ably calculated. Therefore, the theoretical prediction can be considered to
be clean.
It is interesting to note that the CP violation parameter
(7I'°vvlViKd
'f}1[°VV = (7I'° vv lVi Ks)
as defined in the 271' case has i'f}1[ovvi » iti, although the both final states
have CP = +1.
The current experimental measurement for this branching fraction is
< 5.9 X 10- 7 with 90% confidence by the KTeV experiment [22], which is
still far from the expected number. However, there are several proposals to
observe the decays in the near future.
4. B-meson System
4.1. THE STANDARD MODEL DESCRIPTION
4.1.1. Some Elements of The CKM Matrix

Among the nine elements of the CKM matrix, five of them related to the
third generation play important roles in the B meson system: vtd, Vub, vts,
Veb and vtb' In the approximation given in equation 27, the phases of the
five elements are given by
where
(PI = tan- 1 -1 'f} ,<P3 = tan- i "1,~<P3 = tan- i )..2'f}.
-p P
Figure 7 shows the angles in p and 'f} planes. Note that <Pi and <P3 are often
referred to as (3 and "(. Clearly ~<P3 is very small, '" 0.02.
4.1.2. Oscillation Amplitude

In the Standard Model, B-B oscillation is totally governed by the short
range interactions, i.e. the box diagrams. Furthermore, only the top quark
plays a role in the box diagram due to the large top quark mass (see equation
26) and the structure of the CKM matrix;
1m
----~~~~----~------~--~--~Re
o !p 1
I
P(l-A,z/2)
Figure 7. Three elements of the CKM matrix, vtd, Vub, and vts and the definitions of
<PI , <P3
and 8<p3.
as seen from equation 27.

Therefore, the off diagonal element of the mass matrix, M12 IS gIven
by [20]
G~f~dBBdmBdm?,v * 2
M12 = - 61f2 1]BdS(Xt)(~d vtb) for Bd (28)
where fBd' BBd and mBd are the decay constant, B-parameter and the mass
of the Bd meson.
Similarly for the Bs meson, we obtain
G~f~ BBsmBsm?,v * 2
M12 = - s 61f2 1]BsS(Xt)(~s vtb) for Bs
where fBs' BBs and mBs are the decay constant, B-parameter and the mass
of the Bs meson.
The phase of M12 is then given by
for Bd
The parameter n2
can also be determined by taking the absorptive
part of the box diagrams with charm and up quarks in the loops. The
phase difference between M12 and F12 is given by
1
argM12 - argF12 = 1f + '38 (mc)2 x {
mb 1] (29)
192 T.NAKADA
i.e. sin(arg M12 - arg r 12 ) is small for Bd and very small for Bs. Note that
M12 and n2 are antiparallel. Therefore, the approximations for ( , m± and
r± given on page 179 are valid with n = 1. Since we will rely on the
Standard Model description of M 12, and our experimental knowledge of
the decay amplitudes is still limited, we adopt b) <PM base. We refer the
mass eigenstate with larger mass as Bh (B-heavy) and the other Bl (B-light)
with their masses and decay width are given by:
and
ml = M -IMd, rl = r + Ird
respectively, and Bh (Bl) corresponds to P + (P _) defined in equation 13.
For both Bd and Bs, we can now derive
tlr = I r121 ~ 37rm~ ~ 5 x 10- 3 for Bd and Bs (30)

tlm M12 2m~S(xt)
for mb = 4.25 GeV, mw = 80 GeV and mt = 174 GeV, where tlm and
tlr are defined as positive:
Using the measured values of tlm = (0.464 ± 0.018) x 1012 ns- 1 and the
average lifetime T = 1/1'= (1.54±0.03) x 10- 12 s for the Bd mesons, where
l' is the averaged decay width, it follows that
/j.,r ~ 4 x 10- 3
r
and /j.r can be neglected in the decay time distribution for the Bd system.
For the Bs mesons, using the measured lifetime (1.54 ± 0.07) x 10- 12 s, it
follows that
tlr
- ,- ~ 0.1 for Bs.
r
The effect of tlr is still not large, but can no longer be neglected in the
decay time distributions.
The small decay width differences of the Bd and Bs systems do not allow
to separate one mass-eigenstate from the other, which can be done for the
kaon system by creating a KL beam. Therefore, CP violation cannot be
established by just observing the decays as in the case of KL -7 27r. We
either have to compare the decay rates of the initial BO and initial BO states
or measure the time dependent decay rates of at least one of the two cases,
i.e. either initial BO or BO•
Since b..m = 21Ml2i, one can extract
i.e. p and 'T), from the measured BO_Bo oscillation frequency Llmd using
equation 28. However, theoretical uncertainties in calculating the decay
constant and B-parameter are considerable and limit the accuracy on the
extracted value of Ivtdl 2. If the B~-B~ oscillation frequency Llms = 21Mi21 is
measured, Ivtdl 2 can be determined with much small uncertainty by using
the ratio Llmdl Llms, due to better controlled theoretical errors in fBdl fBs
and BBd I BBs· However, the frequency of the B~ -B~ oscillation is expected
to be > 1/).2 = 20 times larger than that of the BO_Bo oscillation and we
still have to wait for sometime before it is measured.
Since IM121 rd « 1, ( given by equation 19 can be further approxi-
mated as
(~ [1- ~ ~ (;;12J] e-i~M (31)
where 'PM = arg M12 as before. Seen from equation 29 and 30, the approx-
imation 1(1 ~ 1 is accurate to 10- 3 or better.
Similar to the kaon system, CP violation (and T violation) in the oscil-
lation can be measured from the time-dependent rate asymmetry between
the initial B O decaying into semileptonic final states with e+ or J.l+, R+(t)
and the initial B O decaying into semileptonic final states with e- or J.l-,
R-(t). The asymmetry is given by
which is a very small signal.

From now on, we assume
for both Bd and Bs and b..r = 0 for Bd.

In summary, the two mass eigenstates are given by
IBh; ~ [IB; + e-i~MIB;]

IBI ; ~ [I B) - e- i ~M IB;]
and
194 T.NAKADA
for Bd and Bs. For the decay width, we have

Fl = Fh for Bd
Fl = Fo + IFd, Fh = Fo -IF121, b..r = rl - rh for Bs
4.1.3. Time Dependent Decay Rates

Since b..F is small in the B meson system, it is more convenient to derive the
time dependent decay rate from the particle-antiparticle base rather than
the mass eigenstate base. Using, equations 10 and 14 the time dependent
decay rates for the final state f can be derived as
Rr(t) (32)
Rr(t) (33)
where l' is the averaged decay time, l' = (F+ + F_) /2, and Ar is the instan-
taneous decay amplitude for the pO -7 f decays. The two time dependent
functions, I+(t) and L(t), are given by
b..F . b..F
(1 + ILrI2) cosh 2 t + 2RLr smh 2 t
(1 - ILrI2) cos b..m t + 2~Lr sin b..m t .
The parameter Lr is given by
Lr = (Ar
Ar
where Ar is the instantaneous decay amplitude for the pO -7 f decays.
The time dependent decay rate for the CP conjugated final states f CP
are derived to be
- 2 _
Rrcp(t) ex IAr~pl e- rt [1~P (t) + T},P (t)] (34)
Rccp(t) ex IArcptl(12 e- rt [1~P (t) -1~P (t)] (35)
where Arcp is the instantaneous decay amplitude for the pO -7 f CP decays.

Two time dependent decay rates, ~P (t) and I~P (t) are given by
I~P (t) = (1 + ILf P I2) cosh ~ t + 2RLfP sinh b..; t

T},p(t) = (1-ILfPI2)cosb..mt+2~LfPsinb..mt
where the parameter, LfP, is given by

LfP = ~ AfcP
( AfcP
and AfCP is the instantaneous decay amplitude for the pO -t f ep decays.
The decay rates Rf(t) and RfCp(t) are CP conjugate to each other and
so are R f (t) and Rrcp (t). If there exists any difference between the CP
conjugated processes, this is a clear sign of CP violation.
The final state f can be classified into the following four different cases:
1. Flavour specific final state (Af = AfcP = 0 or AfcP = Af = 0)
II. Flavour non specific final state
II-a. CP eigenstate (Ar = AfCP and Af = AfCP)
II-b. mixed CP eigenstate (Af = Arcp and Ar = AfcP)
II-c. CP non eigenstate
4.1.4. CP Violation: Clean Case

The contribution to the BO decaying into J j 'ljJ Ks is dominated by the tree
diagram with Veb Ves· Although there exist some contribution from the pen-
guin diagrams, the dominant penguin diagram contribution has the CKM
phase v.:b vts which is close to that of the tree diagram (Figure 8). Thus ,
we can safely assume that there is no CP violation in the decay amplitude
and the ratio of the B O and B O decay amplitudes is given only by the CKM
part. By noting that CP(Jj'ljJKs) = -1 we obtain
A(130 -t J j'ljJ Ks) (Veb Ves Vu*s Vud)2
A(BO -t J j'ljJ Ks) Web Ves V;s Vudl 2
Using the formulae developed in the previous section, the time dependent
rates for the initial B O decaying into Jj'ljJKs, RJj'ljJKs(t), and that for 130
decaying into J j 'ljJ Ks, R J j'IjJ Ks (t) are given by
RJj'ljJKs (t) ex (1 + <;S LJj'Ij;Ks sin 6.m t)

e-i't
RJj'ljJKs (t) ex e-i't (1 - <;S LJj'ljJKs sin 6.m t)
which allow to extract
~ _ ~ ( A(130 -t J j'ljJ Ks)) _ ~ [(v.:'d vtb V';;b Vcs Vd's VUd)2]

~LJj'ljJKs - ~ (x A(BO -t Jj'ljJKs) - -~ 1~'dvtbV';;bVcSV;sVudI2
With the Wolfenstein parameterization, it follows that
<;S LJj'ljJKs = - sin2<Pl .

196 T.NAKADA
B
d, S d, S d, d,
•
S S
•
Figure 8. Tree and penguin diagrams contributing to the BO ~ J N Ks and B~ ~ J N ¢i
decays.
The same argument holds for the Bs -+ J j 1f; cp decays and from the time
dependent decay rates
RJ/,/)(p(t) <X e-i't (COSh ~rt+2!RLJN4>Sinh ~rt+~LJN4>Sin~mt)

-RJN4>(t) <X e- i't (~r ~r
cosh Tt + 2!RL J N4>smh Tt -
. C'l<
';,,5,
•
LJN4>sm~m t
)
one can extract
~LJN4> = ~ [( x ~~!~: ~;~:~l = -sin2oCP3

Note that we assumed in the calculation above that C P (J / 1f; cp) = + 1,
i.e. the J /1f; cP state is in the lowest orbital angular momentum state of
l = O. If there exists the l = 1 state with CP(Jj1f;cp) = -1, the measured
~ LJN4> will be diluted and the fraction of the CP = -1 state must be
experimentally measured. If there is the same amount of C P = + 1 state
and CP = -1 state, ~ LJN4> will vanish.
An even cleaner decay channel is BO -+ D*'f7l'±. There is only one tree
diagram, b -+ c + W+ followed by W+ -+ u + d, which contributes to the
BO -+ D*-7l'+ decays. The same final state can be produced from the BO
decays with another tree diagram, b -+ u + W- followed by W- -+ c + d
(Figure 9). Therefore, the time dependent rate for the initial BO decaying
into D*-7l'+ is given by
RD·-(t) <X e
-i't [ (1 -ILD'-1l'+12)
1 + (1 + ILD'-1l'+12) cos~mt +
2~LD'-1l'+
(1 + ILD'-1l'+12)
'
sm~mt
1
where
~:../c~
B
b
B
b
.,' . u
d d dOIl( d
Figure 9. Tree diagrams contributing for the B O -T D*-7r+ and BO -T D*-7r+ decays.
The weak phase of A(BO -t D*-7r+) is given by VubVcd and that of A(B°-t
D*-7r+) by Vcb Vud· The phase of L D *-7r+ is then derived to be
arg Vub - arg M12 + <{Js

-4>3 + 24>1 + <{Js
where <(Js is a possible strong phase difference between the b -t u + W-
and b -t c + W+ tree diagrams.
CP conjugated decay amplitudes of A(B"° -t D*-7r+) and A(BO -t
D*-7r+), i.e. A(BO -t D*+7r-) and A(BO -t D*+7r-) respectively, are ob-
tained by taking the complex conjugate of the weak amplitudes while the
strong phase remains unchanged. Thus for D*+ 7r- we obtain
where
LC: _ = ~ x A(BO -t D*+7r-)
D 7r+ ( A(BO -t D*+7r-)
and the phase of Lg:- 7r + is given by
- arg Vub + arg M12 + <{JS

4>3 - 24>1 + <{Js
From the two time dependent decay rates, we can extract 4>3 - 24>1.
Note that
i.e. the effect we have to measure is small.

198 T.NAKADA
The CP conjugated time dependent decay rate distributions are given

by
and
which can be used to obtain the same information.

A similar method can be used for the B~ -+ DfK± decays to extract
(P3 - 20rP3· The effect is larger since
4.1.5. CP Violation: Not So Clean Case

The penguin contribution to the Bd -+ 1l'+1l'- decay was originally thought
to be small and the decay would be dominated by the b -+ u + W tree
diagram. However, the discovery of B(Bd -+ K±1l''f) > B(Bd -+ 1l'+1l'-)
indicates that the contribution of the penguin diagrams to the Bd -+ 1l'+1l'-
amplitude should be '" 20% or more.
Due to the penguin contribution, the phase of the BO -+ 1l'+1l'- decay
amplitude deviates from that of V;b ' Furthermore, CP violation in the decay
amplitude could be present. Evaluation of those effects involves calculating
contributions from different diagrams accurately. Strong interactions may
play an important role as well. Therefore, this decay mode may not be ideal
to make precise determinations of p and .,., from CP violation.
4.2. CASE WITH NEW PHYSICS
Decay processes where only the tree diagrams contribute should be unaf-
fected by the presence of physics beyond the Standard Model. Therefore,
Webi and Wubi obtained from the semileptonic decays of B mesons would
not be affected by the new physics and A and p2 +.,.,2 can be obtained even
if physics beyond the Standard Model is present.
New physics could generate BO-Bo and B~-B~ oscillations by new par-
ticles generating new box diagrams. They could also generate a tree level
flavour changing neutral current contributing to the oscillation. Since these
contributions are through "virtual" states, they contribute to M12 with

little effect on r 12 , i.e.
M12 = M 12SM + MNP

12,
r 12 = r 12SM
where Mr2M and rr2M are due to the Standard Model and Mf!l is the
contribution from the new physics. The measured !:lm is given by 21Md
and can no longer used to extract Ivtdl 2 due to Mlit.
Since
r12\
\M12 = 21 rfr 1
!:lm
remains small, CP violation in the oscillation remains small as seen from
equation 31. Therefore,
is still valid. However, note that
'PM == arg M12 =I arg Mfr·

Decay amplitudes from the penguin diagrams can be affected by physics
beyond the Standard Model since new particles can contribute virtually in
the loop. Therefore, the modes such as Bd decaying into 11"+11"-, K±1I"'f may
have some contribution from the new physics.
Since the decays Bd -+ J / 't/J Ks and Bs -+ J / 't/J ¢ are tree dominated,
they are little affected by new physics. Therefore we have
_A-,-(B-::-O_-+_J...:..../-,-'t/J_K-,-s) = _ A (B~ -+ J / 't/J ¢) = -1

A(BO -+ J/'t/JKs) A(B~ -+ Jj't/J¢)
with the phase convention due to the Wolfenstein parameterization and
and studies of the time dependent decay rates give arg M 12 .

The Bd -+ D*1I" and Bs -+ DsK decays are generated by only the tree
diagrams and not affected by new physics. Therefore we have
and
arg LD*+n- = ¢3 + arg M12 + 'Ps
and studies of the time dependent decay rates provide arg M12 + ¢3. Simi-
larly studies can be done for Bs -+ DsK.
200 T.NAKADA
By combining the measurements of Bd -+ J11/J Ks and -+ D*7r or Bs -+

J 11/J cP and -+ DsK, the angle cP3 can be determined even with presence of
physics beyond the Standard Model. By comparing the result from Bd and
that from Bs , consistency of the method can be tested. Since the phase
of Vub is given by cP3 and its modulus is measured from the semileptonic
decay, p and "7 can be extracted. Once A, A, p and "7 are known, Mi'r can
be calculated and from the measured D..m and arg M 12 , the new physics
contribution Mfft is obtained. This can be used to identify the nature of
the new physics contributing to the oscillation.
4.3. EXPERIMENTAL PROSPECTS
A possible experimental programme for the study of CP violation in the B

meson system and search for physics beyond the Standard Model can be
summarised in the following steps:
1. Determination of IVcb I and IVub I from semileptonic (and some hadronic)
decays.
2. Measurement of D..m for Bd and Bs
3. Measurement of '25LJ/'rf;Ks
4. Measurement of LJ/'rf;¢Jl LD*h± and LD~K±
The first step has been made by ARGUS and CLEO at 1(4S) machines
and the four LEP experiments. BABAR and BELLE at the new asymmetric
1 (4S) machines and CLEO will improve the precisions on those determi-
nations. Future improvement of theory is also an important factor. Half of
the second step, D..m(Bd) was done by ARGUS, CLEO, UA1 at the SPS
Collider, the four LEP experiments, SLD at SLC and CDF at the Tevatron.
For D..m(Bs), we may have to wait for the next data taking by CDF, DO
and HERA-B. The third step will be made by BABAR, BELLE, CDF, DO
and possibly HERA-B by the year 2005.
After the second step, four parameters of the CKM matrix are all defined
within the framework of the Standard Model, e.g. A, A, p and "7. The
third step provides an additional information tan- 1 "7/(1 - p) within the
framework of the Standard Model and consistency of the CKM picture can
now be tested.
As demonstrated in the previous chapter, if physics beyond the Standard
Model exists, the fourth step is needed to clearly establish the evidence of
new physics and separate the effect due to the Standard Model and that
from new physics. After the third step, only p2 + "7 2 will be known from
lVubl and the information on tan- 1 "7/(1- p) is spoiled by new physics. Only
after the fourth step, p and "7 can be determined, together with isolating
the new physics contribution.
For the last step, new generation of experiments with statistics much
higher than 1010 B mesons are needed. The Bs meson is an essential in-
gredient. After 2005, LHC will be the most powerful source of B mesons.
Experiments must be equipped with a trigger efficient for hadronic decay
modes to gain high statistics for the necessary final states. Particle identi-
fication is also crucial in order to reduce background. LHCb is a detector
at the LHC optimised for CP violation studies with B mesons. The two
general purpose LHC detectors, ATLAS and CMS can contribute only to
a limited aspect of the fourth step. A proposed experiment at Tevatron,
BTeV, can also make the last two steps.
Clearly CP violation is expected in many other decay channels. For
many of them, there are some theoretical problems for making accurate
predictions. However, they can be used to make a systematic study which
will provide a global picture whether CP violation can fit into the CKM pic-
ture. With all those experiments , we continue to improve our understanding
of CP violation and hope to discover physics beyond the Standard Model.
Acknowledgement
The author is very grateful to the organizers of this school for their ex-
tended hospitality and efforts to prepare such a stimulating environment.
R. Forty is acknowledged for reading this manuscript and giving many use-
ful comments.
References
1. C. S. Wu et al., Phys. Rev. 105 (1957) 1413.
2. R.L Garwin et al., Phys. Rev. 105 (1957) 1415,
J. I. Friedman and V. L. Telegdi, Phys . Rev. 105 (1957) 1681.
3. M. Gell-Mann and R. P. Feynmann, Phys. Rev. 109 (1958) 193,
E . C. G. Sudarshan and R. E. Marshak , Phys. Rev. 109 (1958) 1860,
J. Sakurai, Nuovo Cim. 7 (1958) 649 .
4. J. H. Christenson et al., Phys Rev. Lett. 13 (1964) 138.
5. N. Cabibbo, Phys. Rev. Lett. 10 (1963) 531,
M. Kobayashi and K. Maskawa, Prog. Theor. Phys. 49 (1972) 282 .
6. A.D. Sakharov, JETP Lett. 6 (1967) 21.
7. See for example M.B. Gavela et al., Modern Phys. Lett. 9A (1994) 795.
8. V. Weisskopf and E. Wigner , Z. fur Physik , 63 (1930) 54.
9. For pioneering works see
A. Pais and S. B. Treiman, Phys. Rev. D12 (1975) 2744,
1. B. Okun et ai , Lett . Nuovo Cimento 13 (1975) 218,
M. Bander et ai , Phys. Rev. Lett. 43 (1979) 242,
A. B. Carter and A. I. Sanda, Phys. Rev. D23 (1981) 1567,
I. I. Bigi and A. I. Sanda, Nucl. Phys. B193 (1981) 85 .
10. More details can be found in
T. Nakada, CP Violation in K- and B-Meson Decays, PSI-PR-91-02, 1991.
11. G . Liiders, Dan. Mat. Fys. Medd. 28 , No5 (1954) ,
W. Pauli, In Niels Bohr and the development of physics, ed . W. Pauli, pp. 30, New
York, Pergamon Press, 1995,
202 T.NAKADA
R. Jost, Helv. Phys. Acta, 30 (1975) 409,

G. Liider, Annals of Physics 2 (1957) 1.
12. C. Caso et ai, The European Physical Journal C3 (1998) 1
and 1999 off-year partial update for the 2000 edition available on the PDG WWW
pages (URL: http://pdg.lbl.govl).
13. A. Angelopoulos et al. [CPLEAR Collaboration), Phys. Lett. B444 (1998) 43.
14. A. Apostolakis et al. [CPLEAR Collaboration], Phys. Lett. B458 (1999) 545.
15. W . Ochs, MPI-Ph/Ph 91-35, (1991).
16. G. D. Barr et al. [NA31 Collaboration], system," Phys. Lett. B31T (1993) 233.
17. L. K. Gibbons et al., epsilon)," Phys. Rev. Lett. TO (1993) 1203.
18. A. Alavi-Harati et al. [KTeV Collaboration], Phys. Rev. Lett. 83 (1999) 22.
19. V. Fanti et al. [NA48 Collaboration], neutral kaon," Phys. Lett. B465 (1999) 335.
20. For details of the Standard Model description of the K and B system and further
references, see the following articles:
A. J. Buras, Weak Hamiltonian, CP violation and rare decays , hep-ph/9806471,
A. J. Buras, CP violation and rare decays of K and B mesons, hep-ph/9905437 .
21. L. Woifenstein, Phys. Rev. Lett. 51 (1983) 1945.
22. A. Alavi-Harati et al. [The E799-II/KTeV Collaboration], gamma," hep-ex/9907014.
HERA - EXPERIMENTAL RESULTS
R.J. CASHMORE
CERN
CH-1211 Geneva, Switzerland
1. Introduction
In these lectures I will concentrate on the deep inelastic scattering ob-

served at HERA [1]. HERA, which collides either electrons or positrons
(of rv 27.5 GeV) with protons (of 820 or 920 GeV) is designed to study the
structure of the proton and its consistuents the quarks as well as probe elec-
troweak interactions and search for exotic phenomena. At the same time
extensive studies can be made in photoproduction and in the final states
in deep inelastic scattering but these topics will not be dealt with in this
paper.
From 1994 to 1997 experiments were conducted on e+p scattering with a
proton beam of 820 GeV while in 1998 and 1999 e-p scattering was studied
with a proton beam of 920 GeV. The goal is to reach as high a Q2 in the
scattering process as possible so that the 'resolving' power for any structure
is as fine as possible.
The measurements are made in the two general purpose detectors H1
[2] and ZEUS [3]. H1 uses tracking chambers and a liquid argon calorime-
ter while ZEUS also has tracking but a compensating scintillator depleted
uranium calorimeter. Both detectors are hermetic so that in deep inelastic
charged current interactions the presence of a neutrino can be inferred by
the energy and momentum imbalance produced by its lack of interaction.
These lectures will mainly deal with the e+p scattering where H1 has
recorded 37pb- 1 and ZEUS 47pb- 1 .
2. Structure and Structure Functions

The cross-section for deep inelastic scattering can be written as
d 2 a(l±N) 2 2 2 2 IN 2]
dxdQ2 = 21Ta
xQ4
[ IN IN
. Y+F2 (x, Q ) - Y- xF3 (x, Q ) - Y FL (x, Q) (1)
203
204 R.J. CASHMORE
with Y± = 1 ± (1 - y)2 where
-l = sxy (2)
P.q
Pk (3)
and the variables are indicated in Fig. 1.
k
q:
:,,(, w, z
I
Figure 1. Diagram of DIS interaction.
F2, F3 and FL are structure functions which have simple interpretations

in the quark model.
In particular
(4)
and XF3 (related to the ZO exchange) is negligible for Q2 « Mly also

has a simple interpretation in terms of cos 8*, the lepton quark scattering
angle
1- cos 8* 8*
y = - - -2- = s i n2 - 2 (5)
In studying the structure of the proton at comparatively low Q2 F3

and FL are treated as QeD corrections. The results [4, 5] of the F2(x, Q2)
extractions are shown in Fig. 2 and Fig. 3, which include lower energy
results. In Fig. 2 a rapid rise is observed in F2 at low x, much larger than
HERA - EXPERIMENTAL RESULTS 205
expected from fixed target results. This corresponds to much larger gluon
and quark parton distributions at low x than was originally expected. If
scaling had been corrected F2 plotted against Q2 at a given x would have
been flat. In Fig. 3 we see the scaling violations clearly demonstrated with
the increases at low x and the depletion at high x as a function of Q2.
ZEUS Preliminary 1996-97

2 r-------~----~~------_. 2
~ ~
'..
t. Ii·" Ii="
Ql:::J5
t. I.
\ '...
".'. ",
"
".......
2
""'". ''' . .\
Q1",70
t
~\ rio i
....0,
;.
.,-. .....
....,. .'."
cr= 12.
.'
cr= •• "
2 '"
r.•• ~
.... 11>
t.
,~.
. ...... .,et "",.
,.
Ql=150 cr=21JO cr=2S"
2
'~
!
1\
cr= '5.
fro\.
' ,. cr= "0 .'" cr= ".
",...,
O~~~~~~L-~~~~~~
I.' ,,,.. , ,' II" I.' Ir II" I. ' III ' 111' /lr' 'If' Iff.! 'tT' /11 ' I
o ~ ~ ~' ~ Wi ~ ~ ~ ~ w' ~ ~ ~ ~ ~ I
X X
Figure 2. F2 as a function of x for different Q2.
These values of F2 can be analysed with the DGLAP formalism [6] which
describes QCD evolution of the structure functions. Excellent fits [7] can
be obtained which provide gluon, g(x,Q 2 ), and quark, q(x,Q 2 ) distribution
functions. Such a distribution at Q2 = 20 Ge V 2 is shown in Fig. 4 indicating
the rise at low x.
This large value for the gluon distribution has been confirmed [8] by
studying photon gluon fusion processes which lead to charm production
,+g--7C+C
and the charm quarks are observed by the measurement of the D mesons .
Finally measurements [9] by HI have indicated the consistency of Fdx, Q2)
with QCD expectations.
206 R.J. CASHMORE
s" &8E--~
.-3.iIE-05
.·S.lE-05
t: A:8 (;-- ' -- - .-0.000102
~ A:gc--. -----.~·. .-0.000181 o ZEUS 96-97 Proliminory
&g C---,-~---. --._.. IC-O.OOO2~J • BCDM5.f.r:'65.NMC
A:8[_--·····~·· •
&8 [_._~ •.•.•.••• :. u . .-0.0005
&gc--- -------··· ·_ ,._+---" x-O,oooeJ2

A:g r:.-.. -..~~ ••,• •... .
&8r;: _._ •• A . . .• . . :.! ' .. ,
&R ~ ____ - --_.......... ___a _• ..e.!-- - .·0.0013
&8 [ ____ --.-------..... ..
c- .. ..........* .~
.
. .... . .
A:8 .~ ~ ~~
&8 [:--___ ~-.--......•.•---- ___aV-.- - ,,·0,0021
.-0.00253
A:8 ~=_.-~ ~ ~ ! - .~--
.·0.0032
1.:8 [ ............. ------ ....--------
6~8 ~:~:-==
o:g
_______.
o.g [ - - _a-.........____...... -'_L..~---'-
_.~
__.....-. ______ ..... ____ ._.....__
~...-
J _ ,a_AJ. L L L ___ ~ __
LII. _______________ ___

.-0.005
.-0.001
.·0.013
A;gL-'_ ..... ~ ·~·............ ·~··...........-...· .-0.021
oi ~_..----.___.--.-.--------......----..-------~.-+-- x-0.032
Aj ~ _______.-....__ •__... ,. _______ L..a_..e__*_. . . . _ - - - - - - - . . _ . . . _ _ _ _ x.O.05
6:8 ~____ • •• ... .... ___-~------....... ... • • .-0.08

A·8 •
6:g. ~~--------------..-~....
~ ___ ........................._______________________.-.-.-.• ___ .." __________ x-013
. x-0.18 ]
O:g~ ___ -_____••••••__ _ ' - - - - - '-----*-+-I x.O.28 ] .. .'
A:gc:...._. ... _......~- ...- .••-...... ~ -0.32 J
- -- _._
+ .. . •
A:gC.
--
.-0.'5 J
6:g [" ~.--.---------••--••--- - ---- ----.-----.-------------··------:-.0.55 J
... ..........----------- .
1d 1do 1dooo 1doooo
Ci (GeV')
Figure 3. F2 as a function of Q2 for different x.
3. High Q2 studies - Electroweak Interactions and searches for

Exotic Phenomena
3.1. ELECTROWEAK INTERACTIONS
In these studies, we are interested in both the Neutral Current (NC) cross-
sections mediated by "y and ZO exchange and the Charged Current (CC)
cross-sections mediated by W± exchange. The leading order diagrams, at
quark level, are shown in Fig. 5. The corresponding NC cross-section is
given below
(6)
,,-... 35
~
'Eli .. NLO QeD fit Q2 = 20 GeV 2
~
a.(M!)
30
o Hl1994
(preliminary)
(0.118±0.00S
m,= 1.3 • 1.8 C.
25 ZEUS 1994 (0.113±0.OOS

(preliminary) m,=103 · 1.5 Ce
~ NMC (0.113)
20
ZEUS+Hl 94(0.118)
MRSRI (0.113)
15
CTEQ4M (0.116)
GRV94-HO (0.110)
10
o .4
1
10
"
Figure 4. Gluon density xg{x) as extracted from the DGLAP fit to the F2 data.
2
± (-) I
d
L
(je q ..... v"q
--d:-Q-=-:2::--- ex IVqqf \2
qf
Figure 5. Leading order Feynman diagrams of the neutral and charged current interac-
tion.
208 R.J. CASHMORE
Y± = 1 ± (1 _ y)2
;=fc = E q=d,u,s,c,b A·[+-]
q q q
;=fC = Eq=d,u,s,c,bBq . [q - ill
Aq Q~ - 2Q q v e v q . Pz + (v~ + a~)(v~ + a~) . pi
- 2Q q a e a q . Pz + 4VeaeVqaq . pi
Q2
Pz = Q2+M~
and the CC cross-sections are
d2(J'e+p-tiJX
dxdQ2 =
G2
2:
Pa, [(u + c) + (1 - y)2(d + s)]
d2(J'e-p-tvX
dxdQ2
G2
2:
Pa, [(u + c) + (1- y)2(d + s)] (7)
M2W
Pw =
Q2 + Ma,
The need for ZO exchange is demonstrated in Fig. 6 and recent mea-
surements [10] of the e-p NC cross-section at Q2 > 3000 Gey2 are larger
than the corresponding e+p cross-sections demonstrating the existence of a
positive ,Zo interference in the e-p cross-section. In Fig. 7 we see a com-
parison of the e+p Charged and Neutral Current cross-sections. At high Q2,
where the Z and W exchange are important, the cross-sections are similar
which is a graphic demonstration of the electroweak unification.
In Fig. 8 the Charged Current cross-sections for e+p and e-p scattering,
as measured by ZEUS [11], are compared. HI have obtained similar results.
The e-p cross-sections are an order of magnitude larger than e+p which is
due to the fact that
(u + c) (8)
while
(9)
and y is large at high Q2, i.e. these cross-sections are very sensitive to the
different quark flavours at high x.
Finally it is possible to make a measurement of the W mass from the
variations of the cross-sections with Q2. Fits lead to the value of
(1O)
which is consistent with the Particle Data Group value of Mw = 80.4 ±
0.10 Ge Y, and complements the e+ e- and pp measurements of the 'timelike'
W mass.
ZEUS NC 1994 - 97
• •• L t T······ · ··········
___ to
· · · · · · · · · . ~:::~b)
.c
,eo 9 • ZEUS 94-97 e+p NC
~ 8
t5 7
t 6
5
4
3
Q2>10000 GeV 2
O.t 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 t

X
Figure 6. Measured NC cross section compared to the SM prediction for three assump-
tions of the ZO masses.
3.2. SEARCH FOR EXOTIC EFFECTS
There was a lot of excitement generated in rv 1996 by the observation in

ZEUS and HI of an apparent excess of events at high Q2 [12, 13]. HI implied
that this might be due to the existence of an eq resonance (a leptoquark)
of mass rv 200 GeV, although the ZEUS excess was not concentrated in
this region. Subsequent data however has not substantiated this result and
exclusion zones for MLQ and f3(LQ ~ eq) are now comparable to those
derived from the Tevatron where leptoquarks would be pair produced.
The Deep inelastic scattering can also be used to set limits on new
lepton-quark contact interactions, perhaps indicative of new higher mass
physics processes e.g. new Z' bosons, leptoquarks. These appear as terms
ex:b in the Lagrangian and lead to different cross-sections at high Q2.
The results obtained from HERA data [14] depend on the exact form of
the contact interaction but lead to lower limits for the A's in the range
1.5 - 5.2 TeV comparable to results obtained from CDF [15] and the LEP
experiments [16].
210 R.J. CASHMORE
HERA e+p DIS cross section 94 - 97

Preliminary
10
.,
o
N
neutral current
"0
'0
"C
·4
10
·5
10
·6
10
• ZEUS
·7
10 • H1 y< 0.9
10 3
Figure 7. Comparison of the measured CC and NC cross section with the SM predictions.
Finally it is important to look for DIS events which lead to leptons of

different sign or flavour, as these would again indicate new physics. Such
events would occur, at a low level, due to direct W production (with its
subsequent decay to J.W, or ev). HI have reported an excess of J..L production
events in e+p scattering while ZEUS observes a number consist ant with W
production. Only time and more data will resolve this issue.
4. Outlook
This brief paper has dealt with deep inelastic scattering, the main raison
d'etre for HERA. However extensive studies have been made of diffraction,
final states and photoproduction leading to a very full view of QeD at
HERA [17J . There is still much to be done in these areas.
However the really important changes at HERA will be the luminosity
upgrade in 2000 (leading to > 4 increase) and the introduction of polari-
sation rotators for the lepton beams. The increased luminosity will further
the studies at high Q2 and allow the resolution of some of the outstanding
anomalies while the rotators will allow an in-depth study of the electroweak
interactions, particularly in the light (u and d) quark sector.

.....>
·1
10
• e" p Data
"
~
... e+pData
..... ·2
,Q 10
S.
...
0' ·3
'C
'0 10
'C
·4
10 CTEQ 40" p Ep = 920 GeY
CTEQ 40" p Ep = 820 GeY
·s CTEQ 40 e' p Ep = 820 GeY
10
·6
10
10
., -{ nsmte8YM
·8
10
10 4
Q2 (GeV2)
Figure 8. Comparison of CC cross section for e+p and e-p scattering.
HERA is an essential machine in the study of the proton structure. This

is an important goal in its own right but is absolutely essential if we are
to study pp interactions at high energies (the LHC) with any quantitative
reliability.
References
1. 'HERA, A proposal for a Large Electron-Proton Colliding
Beam Facility at DESY', DESY HERA 81-10;
G.-A. Voss and B.H. Wiik, 'The electron-proton collider
HERA', Ann. Rev. Nucl. Part. Sci. 44 (1994), 413-452.
2. HI Collaboration, I. Abt et al., Nucl. Instrum. Methods A 386, 310 and 348 (1997)
3. ZEUS Collaboration, Status of the Detector
Status Report 1993 - The Blue Book, ZEUS 93, DESY (1993) .
4. ZEUS-F2 (1994) , ZEUS Collab. , M. Derrick et al. , Z. Phys . C 72 , 339 (1996);
ZEUS-F2 (1994), ZEUS Collab., M. Derrick et al., Z. Phys. C 69, 607 (1996) ;
Measurement of the Proton Structure Function F2 in e+p Collisions at HERA,
Submitted paper to the XXIX International Conference on High Energy Physics,
Vancouver , July 23-29, 1998.
5. HI Collaboration, A. Aid et al. , 'A Measurement of the Proton Structure Function
F 2(x , Q2) ' ,Nucl. Phys. B 470, 3 (1996);
HI Collaboration, C. Adloff et al., 'A Measurement of the Proton Structure Function
F2(X, Q2) at low x and low Q2 at HERA', Nucl. Phys. B 497, 3 (1997) ;
212 R.J. CASHMORE
'Precision Measurement of the Inclusive Deep Inelastic ep Scattering Cross Section

at Low Q2 at HERA', Submitted paper to the XXIX International Conference on
High Energy Physics, Vancouver, July 23-29, 1998.
6. G. Parisi, Proc. 11th Rencontre de Moriond, ed. J. Tran Thanhm, ed. Fronties,
1976; G. Altarelli and G. Parisi, Nucl. Phys. B 126, 298 (1977) . G. Altarelli, Nucl.
Phys. B 81, 1 (1981);
V.N. Gribov and L.N. Lipatov, Sov. J. Nucl. Phys. 15, 438 (1972);
L.N. Lipatov, Sov. J. Nucl . Phys. 20, 96 (1975);
Y.L. Dokshitzer, Sov. Phys. JETP 46, 641 (1977).
7. A.D. Martin et al., Eur. Phys. J. C4 (1998) 463 ,
CTEQ Collaboration, H.L. Lai et ai., hep-ph/9903282 .
8. ZEUS Collaboration, J.Breitweg et al. , 'Measurement of Dd production and the
charm contribution to F2 in deep inelastic scattering at HERA', Eur. Jour. Phys.
C 12, 35 (2000);
HI Collaboration, C. Adloff et al., 'Measurement of D* Meson Cross Sections at
HERA and Determination of the Gluon Density in the Proton using NLO QCD',
Nucl. Phys. B 545, 21 (1999) .
9. HI Collaboration, C. Adloff et al., 'Determination of the Longitudinal Proton Struc-
ture Function FL(X , Q2} at Low x, Phys. Lett. B 393, 452 (1997).
10. HI Collaboration, C. Adloff et al. , 'Measurement of Neutral and Charged Current
Cross-Sections in Positron-Proton Collisions at Large Momentum Transfer', DESY-
99-107, submitted to Eur. Phys. J. C.
11. ZEUS Collaboration; J.Breitweg et al. 'Measurement of High Q2 Charged-Current
e+p Deep Inelastic Scattering Cross Sections at HERA', Eur. Jour. Phys. C ??, ??
(1999).
12. ZEUS Collaboration, J . Breitweg et al. , 'Comparison of ZEUS Data with Standard
Model Predictions for e+p -+ e.f X Scattering at High x and Q2" Z. Phys. C 74,
207 (1997).
13. HI Collaboration, C. Adloff et al., 'Observation of Events at Very High Q2 in ep
Collisions at HERA', Z. Phys. C 74, 191 (1997) .
14. ZEUS Collaboration; J.Breitweg et al., 'Search for Contact Interactions in Deep
Inelastic e+p -+ e+ X Scattering at HERA';
DESY 99-058 (May 1999), submitted to the Eur. Phys. J. C;
HI Collaboration, S. Aid et al., 'Leptoquarks and Compositeness Scales from a
Contact Interaction Analysis of Deep Inelastic e( +-)p Scattering at HERA, Phys.
Lett. B 353, 578 (1995).
15. CDF Collaboration, F. Abe et ai., Phys. Rev. Lett. 79, 2198 (1997)
16. OPAL Collaboration, G. Abbiendi et al. , CERN-EP/98-108 (ICHEP98, Abs.264);
ALEPH Collaboration, ALEPH 98-060 (1998) (ICHEP98, Abs. 906);
L3 Collaboration, M. Acciarri et ai., CERN-EP /98-31 (1998) .
17. A.T . Doyle, talk at the XXIX International Conference on High Energy Physics,
Vancouver, July 23-29, 1998.
RECENT RESULTS AND PERSPECTIVES AT CDF AND DO
H.E. MONTGOMERY
Fermi National Accelerator Laboratory
P.O. Box 500
Batavia, IL60510 U.S.A.
Abstract. Over the course of the past years the experimental measure-
ments performed by the two large collaborations, CDF and D0, at the
Fermilab Tevatron Collider have fueled advances in our understanding of
physics at the energy frontier. At the present time the accelerator complex
and the two detectors are undergoing substantial improvements. In this pa-
per, we provide a discussion of some recent results which in turn provides
a framework within which we can look to future prospects.
1. Introduction
The Tevatron with the aid of its associated detectors, CDF and D0, has
made its most significant mark on experimental particle physics progress
with the observation of the top quark. However this was only one of a
number of important contributions[l, 2]. In this paper we describe some
examples of recent results[3] and look forward to future running of the ex-
periments with significantly increased luminosity. The latter is possible as a
result of the introduction of a new accelerator, The Main Injector, into the
Fermilab complex; this is described in Section 2. Both detectors are under-
going upgrades which will enable them to operate in the new environment
with greatly enhanced capabilities. These changes are briefly described in
Section 3. In Section 4, we discuss the physics accessible at the pp collider
working our way through the physics of the strong interaction, QCD, to the
physics beyond the standard model. At present much of the latter remains
speculative but provides the framework for the future experimental work.
Finally we offer a brief conclusion in Section 5.
213
J.-J. Aubert et al. (eds.), Particle PhYSics: Ideas and Recent Developments, 213-234.
214 H.E. MONTGOMERY
2. The Tevatron and Main Injector

The Tevatron is a pp collider[4]. The energy is 900 GeV in each beam and it
is the highest energy collider in the world. In a hadron collider, the effective
parton-parton energy is controlled not only by the machine energy but also
by the luminosity. At high luminosity the rate of higher momentum scatters
increases. Thus, while tI production was accessible kinematically from the
first day of the collider operation, it was not until ten years later that
sufficient high momentum scatters had been accumulated to make the tI
production observable experimentally.
The luminosity of the machine is, to a good approximation, controlled
by the total number of antiprotons available. In turn this depends on the
production rate and the cooling rate. The new Main Injector, a rapid cy-
cling proton synchrotron operating at 150 Ge V, will be used for antiproton
production. An innovation is the introduction of the Recycler, an 8 GeV
storage ring constructed with permanent magnets. At the end of each Teva-
tron store, approximately half of the initial antiprotons are still present.
Most of the luminosity degradation is a result of beam blow-up. The an-
tiprotons are to be decelerated in the Tevatron, then in the Main Injector
to 8 Ge V, and stored and recooled in the Recycler. After cooling the the
anti-proton capacity of the complex is increased by a factor of two. With
these measures, the instantaneous luminosity in the collider is anticipated
to increase to about 2 x 1032 cm- 2 s- 1 . This will make an integrated lu-
minosity of about 4 fb- 1 available in the next few years perhaps rising to
10- 30 fb- 1 over the next six to seven years. This is to be compared with the
0.1 fb- 1 of the present data set. The Main Injector has been comissioned
and the Tevatron is operational for fixed target physics. At the same time
the energy per beam will be increased from 900 Ge V to close to 1000 Ge V.
For high mass processes, such as top production the cross section increases
by 30-40%.
3. CDF and D0 Detectors

The CDF detector, see Fig. 1, has been operational for more than ten years.
It contains a large solenoid which provides a 1.4 Tesla magnetic field in the
tracking volume. That volume is surrounded by a scintillator/lead/iron
calorimeter and a non-magnetic muon detector. The current upgrade con-
centrates on completely replacing the tracking detectors. The outer tracking
will be provided by a large open cell drift chamber with shorter drift dis-
tances than the previous detector. At inner radii, this is complemented by
a comprehensive silicon detector system both to enhance the tracking capa-
bility and to provide detection of secondary decay vertices. The detection of
B hadrons, both for their own sake and as indicators of the decays of higher
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 215
Figure 1. The upgraded CDF detector.
.
"ONIH f SOt,n .
'
.~.: ..
"J'
;".
'. -;,"J .
.,~:
• '- d
P- ....,l......,_
( ..... ) 0
','
~:.:
':" W-Jl~:::::::"'.,:t:F~~~~~~ ,.
(~)
Figure 2. The upgraded D0 apparatus.
mass states, top and perhaps the Higgs particle, places a strong premium
on this capability.
The D0 detector, see Fig. 2, is characterised by a three-cryostat liquid
Argon/Uranium calorimeter with good electron and jet resolutions. The
muon system consists of detectors inside and outside of large iron toroids
in both central and forward regions. The forward muon system is being
equipped with new trigger and tracking detectors to accomodate the up-
graded accelerator parameters. A new superconducting solenoid has been
installed in the tracking volume and the particle detection will be performed
using a scintillating fiber tracker and a 800,000 channel silicon tracker.
Initially the collider will operate with 496 nsec between collisions of
216 H.E. MONTGOMERY
the bunches but this will eventually be reduced to 132 nsec. Pipelines,
analogue in some cases, digital in others, have been introduced in the front
end electronics in both of the new detectors. The data acquisition systems
have also been upgraded to accomodate event rates of several tens of Hz,
to accomodate the overall luminosity increase. The detectors are to be
operational in early 2001.
4. Physics
4.1. QeD
o 00 Oata 1'1;,,1 < 0.5
"*"JETRAO
CfEQ3M. J.l = 0.5 E;'"
to
I
50 100 150 200 250 300 350 400 450 500
E.r(GeV)
Figure 3. D0 inclusive jets cross section and a comparison with a next-to-leading order
QeD prediction.
Production cross sections in pp collisions are calculated by convoluting

the parton distribution functions in proton and anti-proton, respectively,
with the appropriate hard parton-parton scattering cross section[5]. The
parton distribution functions are derived from a number of measurements,
primarily from lepton scattering, and are evolved to the appropriate hard
scattering scale. That this paradigm works well is demonstrated in Fig. 3
where we see agreement of the calculations with the inclusive jet produc-
tion cross sections from D0 over many orders of magnitude. Deviations
from perfection would be signs of either changes required to the parton
distribution functions or, perhaps, a sign of physics beyond the paradigm.
While there have been alarms, the currently accepted view is that agree-
ment between predictions and data is good[6] . A corollary is that from

these results, from measurements of the dijet mass distributions, and from
measurements of the angular distributions, a lower limit on the scale of any
possible compositeness can be set at about 2-3 TeV.
>'OJ CDF Prelimin ary

()
" - 10
D
C
• O. 1 < 11) < 0. 7
W 1
• 0.7 < 11)1 < 1.4
1.4 < 11) i < 2. 1
"D
"- ..
b
"D - ,
10 ,. 2. 1 < 11) ' < 3. 0
CTEO 4H J
-2
10 MRS DOp
CTEO 3M
-3
10
JETRAD
-4
10
-5
10
·6
10
50 100 . 50 200 250 300 350 400 4 50

[, (GeV)
Figure 4. CDF Inclusive Jets at large rapidity.
The prescription described above should also permit an incisive compar-

ison between cross sections measured at different energies, 1800 Ge V and
630 Ge V, appropriately scaled in transverse momentum. This comparison
should be fairly insensitive to the choice of parton distribution functions;
however what is found is that the theory differs from the preliminary mea-
surement[6 , 7] by about a factor of two. Agreement can only be achieved by
modifying the choices of renormalisation and factorisation scales. Recently
the measurements have been extended[8] in rapidity. Again these measure-
ments are sensitive to different aspects of the parton distribution functions.
In this case, as can be seen in Fig. 4, agreement is good.
Traditionally, experiments at hadron colliders have used a cone algo-
rithm[6] to reconstruct jets. In contrast , work using electron-positron, or
lepton-nucleon collisions has employed algorithms which construct jets from
elementary objects such as the individual charged tracks or the energy de-
posited in a calorimeter by a single hadron. Such algorithms[9] are known
218 H.E. MONTGOMERY
lO .L---L_..L....--J.._.L---J.._.L---'-_~--'--------'
i1 t, ,',' ",:=,'" ",', "C""

~ 0 ~ ~ ~ ~ ~ ~ ~ ~ ~
PTW(GeV/c)
~I
Figure 5. W-boson PT spectrum as measured by D0, the curves are the bounds of the
smeared next-to-leading order predictions.
as kT, Jade, or Durham algorithms. Recently we have started to use them

in pp experiments to look at the jet substructure. D0 has a preliminary
result[lO] which suggests that the multiplicity of sub-jets in gluon jets is
larger than that in quark jets. In the future, this approach could well form
the basis for distinguishing jet identities in other analyses such as Higgs
searches.
When the production of Wand Z bosons was first established at the
CERN SppS there were fewer than ten each of these particles. The number
of W bosons observed by each of CDF and D0 now approaches one hundred
thousand. These data samples have permitted the use of the weakly inter-
acting final state particles in investigations of QeD, in a manner analogous
to the use of virtual bosons in the initial state in neutrino experiments.
Wand Z production is dominated by the lowest order parton model
annihilation of valence quarks from proton and antiproton. The QCD cal-
culations agree very well with the cross sections which have been mea-
sured[ll, 12] with a precision of a few %. The higher order QCD correc-
tions introduce transverse momentum for the bosons and at high PT , the
recoiling hadronic system may contain one or more jets. The transverse
CDF Preliminary
- CDF Data (108 pb") 0.4 Jet Cones

Stat. only Stat. + Syst.
0.065
DYRAD NLO QCD Predictions (a, =Or =Mwl
• MRSA Family c CTEQ4 Family
0.060
0.055
.. ---- ...... .._.,._..... -.... .... . .
·~ -~ ·· o---a ·····
,,_-... .. ____ ....................... -4--,; .... -a'...... ----- ... ---.. -.~- ......-.............. _......-~ ..
0.050 •
0.045
. ---~---- ... -~..--------.- .. - -_... _---_._----------_..
Jet E~1n = 30 GeV
0.040
0.08 0.10 0.14 0.1 6
Figure 6. Fraction of W-boson events with at least one jet as measured by CDF. The
dependence of the prediction on the strong interaction coupling is also shown.
momentum spectra at high momenta are expected to be well described by

perturbative calculations while at low PT non-perturbative effects are ex-
pected. Resummation of some of the logarithms is expected to be necessary.
Recent results on both Z and W spectra[12, 13] are surprisingly well de-
scribed by the extant predictions. This is illustrated for W production in
Fig. 5.
The CDF measurement[14] of the fraction of W production containing
one or more jets above a given threshold, RIO, is shown in Fig. 6. Concep-
tually, this is a classic measurement of the strength of the QeD coupling
strength. The events with at least a single jet contain at least one strong
interaction vertex, the total production is dominated by events with no
such vertex. As usual, since as is not so small, there are higher order
corrections. Nevertheless, the theory at next to leading order provides an
adequate description of the results.
4.2. FLAVOR PHYSICS
Thus far, three generations of quarks have been observed; each generation
contains an up-type quark and a down-type quark. The weak states are
220 H.E. MONTGOMERY
mixtures of the eigenstates of the strong interaction. This mixing[15], is

described by a 3 x 3 matrix of transition amplitudes between the quark
states. The Cabibbo-Kobayashi-Maskawa matrix can be described by four
parameters. One of these is a phase through which the formalism accomo-
dates and describes the CP violation observed in the kaon system. In turn
the parameters of the matrix, assuming that there are only three genera-
tions , can be represented, as shown in Fig. 7, by a triangle. Many of the
properties of this triangle are accessible by measurements of the kaon sys-
tem, however measurements of the properties of B hadrons are becoming
increasingly important.
KO-1t°w
L
1.5 K~-1t°l e
e'/e
11 1
0.5
/ If-If
B-Xd'>j
B-Xd ll
Bd- lf
,,
,,
--------:..'
,
\
\
\
\
B \
-1 1 2
P
Figure 7. The Unitarity Triangle associated with the Cabbibo-Kobayashi-Maskawa
flavor mixing matrix and the measurements possible in the kaon system.
On the scale of 1800 GeV, the mass of the b quark is small; the produc-
tion cross section is about 1/1000 of the total cross section. Measurements
of the cross section by both experiments in the central region are about
a factor of two higher than expected. Production extends over about six
units of rapidity and D0 has made measurements[17] in the forward direc-
tion which are even higher, a factor four, with respect to the predictions.
As a result of the high energy, B hadrons which contain a strange quark
or a charm quark are produced in addition to those containing up and down
quarks. For example, in 1998, CDF observed[18] the Be. This state was ob-
served in a semi-Ieptonic decay mode with a missing neutrino. Nevertheless,
the signal was unequivocal and a good determination of the mass was made.
Using its silicon vertex detector, CDF has accumulated a set of measure-
ments of the lifetimes of various B hadrons. These results are displayed in
Fig. 8. It is immediately clear that, unlike in the charm system, the lifetimes
CDF B Lifetimes
I I
.(So) 1.51 ± 0.05 ps

I-l
.(s+)
I-l 1.66 ± 0.05 ps
.(S~) H 1.36 ± 0.10 ps
"t(A~) ~~- I 1.32 ± 0.17 ps
't(S~) H 0.46 ± 0.17 ps

inc" T(b} ·1 .53 0.04 ps
-r(S+)h(SO) . ........................ .
fel 1.09 ± 0.05
I
0.5 1 1.5
Figure 8. A compilation of B hadron lifetime measurements from CDF.
of charged and neutral B mesons are rather similar. This is understood to

be the result of the dominance of the simplest diagrams as a result of the
high b-quark mass. We also notice that the lifetimes of both the Ab baryon
and of the Bs meson are quite similar to that of the B+ and BO. The life-
time of the Be is affected by the decay of the c quark as well as that of the
b quark, hence the observed factor of two shorter lifetime[19J.
The neutral B mesons mix in a manner similar to the neutral kaons
and the mixing for Bd has been measured. Thus far the attempts to detect
Bs mixing have not met with success. The most recent measurements[20]
from CDF place a limit(~ms < 5.8 ps, Xs = ~ms/r s > 7.9} as good as any
achieved by the LEP experiments or by SLD at SLAC. Extrapolating to
Run II, we expect sensitivities in the range Xs 2:: 25 from each experiment,
likely more from CDF, thus comfortably covering the expected range.
The goal of flavor physics experiments is to measure all the parameters
with sufficient detail to overconstrain the CKM matrix and, if possible, to
break the model. Recently CDF presented results of their measurements of
the CP violating asymmetry in the decay B -t J/'ljJK~ which determines
sin 2,6. They have one sample of events for which the proper decay time is
measured and one for which only the time-integrated measurement is ob-
tained. The results are displayed in Fig. 9. They find[21] sin2,6 = O.79~~::~
222 H.E. MONTGOMERY
--ilm fixed
.......... ilm floating
....
fIl
=N
"O:l
- 2
- ~
1
- ~ ....-
...... -..
.... ~
--_ ....... ......... _.... _.. 1
oW
>- .'
"
, o
.~ ........ ~-,.-
-1 - -1
-2
o 0.05 0.1 0.15 0.2 0.25
ct (em)
Figure g. Measurement of the CP violating Asymmetry in J/1/JK~ from CDF. On the

left the asymmetry is shown as a function of lifetime, and on the right, for an independent
sample, the time-integrated asymmetry is shown.
suggesting a positive value at about the 90% c.l. A feature of this measure-
ment is the use of several different flavor tagging techniques. With approxi-
mately 2 fb- 1 and the upgraded detectors, the uncertainty on sin 2{3 will be
reduced below 0.1 for each experiment. Similar uncertainties are projected
for sin 2a although the interpretation for this case is considered to be more
difficult. Measurement of the third angle, /, will be a challenge.
The mass of the top quark will be considered as an electroweak pa-
rameter and discussed in the following section. As far as the determination
of the couplings and other characteristics of the top quark are concerned,
studies are in their infancy. The observed cross section certainly seems to
be consistent with that expected. Further, the mix of events of different
topologies as yet show little deviation from expectations. These facts limit
somewhat the liberties which can be taken with the coupling vtb between
top and bottom quarks and possible decays, for example into a charged
Higgs boson plus a bottom quark.
One of the fascinating properties of the top quark is its short lifetime.
This implies that the W boson into which it is expected to decay will
have a well defined polarisation. CDF has measured[22] that the fraction
of longitudinally polarised W bosons is 0.97 ± 0.37(stat) ± 0.12(syst); the

expected value is 0.70. The short lifetime means that decays occur before
hadronisation and as a result correlations between the spins of the top
and anti-top quarks, generated by the annihilation production diagram,
are expected to survive and manifest themselves in the relative spins of the
observed states. D0 has made a measurement using the sample of events
in which each of the W bosons decayed into a lepton and neutrino. While
statistics are small, this sample is relatively background free. Along with
that of the b quark, the angular distribution of the charged lepton is the
mOst powerful indicator of this property. The measure of the correlation is
the parameter f'i, defined such that -1 < f'i, < +1 with f'i, = 1 expected in
the standard model. D0 observes[23] f'i, > -0.25 at 68% d.
4.3. ELECTROWEAK PHYSICS
In an earlier section we mentioned the large numbers of Wand Z bosons

produced at the Tevatron. In addition to single bosons, boson pair pro-
duction is also kinematically accessible. The various possible processes are
sensitive to the triple boson couplings. These couplings among the gauge
bosons are a fundamental feature of the non-Abelian electroweak theory
and markedly different from the purely electromagnetic theory in which
the gauge bosons, the photons, carry no charge. In the electroweak theory
the non-Abelian couplings lead to cancellations among the different dia-
grams. For example, without them the production cross section for several
diboson final states would diverge at high energy and would violate unitar-
ity. As a result, searches for these rare processes have led to limits[24] on
the possible deviations from these coupling strengths from their standard
model values. As an indicator of what the future may hold, CDF has ob-
served an excellent candidate for the pair production of Z bosons and D0
has an excellent candidate W Z event.
The copious production of W bosons permits a study of their properties.
The ratio of production cross section times branching ratios for Wand Z
can be related, using measurements from LEP of the Z boson properties, to
the total W-boson width[ll]. The standard model value for the W leptonic
width is used as input as well as the calculated ratio of the production
cross sections. Alternatively the high-transverse-mass tail of the the W
boson event distribution can be used to directly measure [25] the width.
At present the indirect technique is most precise; however, it does depend
on standard model assumptions. The hope would be that in the future
both measurements would be of sufficient precision to provide an additional
constraint on the standard model and a determination of the leptonic width
of the W boson.
224 H.E. MONTGOMERY
The gauge sector of the standard electroweak model is specified by three

quantities usually taken to be the muon weak decay constant , the electro-
magnetic fine-structure constant and the mass of the Z boson. These three
quantities unambiguously lead to a prediction of the mass of the W boson
at lowest order in the theory. Higher order corrections are expected from
loop diagrams containing the fermions and hence dominated by the top
quark, and a diagram with emission and reabsorbtion of the Higgs boson,
if such exists. A precise measurement of the W boson mass is therefore an
important goal.
W bosons are observed by detecting a charged lepton and measuring the
hadronic recoil vector. This permits the neutrino kinematics to be inferred
using momentum conservation, but only in the transverse plane. The mass
of the boson is therefore deduced by fitting templates generated with a
range of masses to one or all of the two lepton transverse momentum spec-
tra and the transverse mass spectrum. (The transverse mass is a quantity
constructed analogously to the effective mass but which is defined using
only the transverse components of the energy-momentum vectors.) Each
of these spectra is sensitive to different aspects of the measurement. The
charged lepton recoils are very sensitive to the W transverse momentum
spectrum but largely independent of the hadron measurements. The neu-
trino recoil is very sensitive to both the boson transverse momentum and
the hadronic measurement. The transverse mass is moderately sensitive to
the lepton and hadronic measurements but relatively insensitive to the bo-
son transverse momentum. Hence, use of all the measurement information
provides powerful cross checks; maximum sensitivity is achieved by combin-
ing all measurements. These features, in addition to the use of the Z-boson
data for in situ calibration, permit the experiments to match reduction
in systematic uncertainties to reductions in statistical uncertainties as the
samples increase.
The current results 80.433 ± 0.079 Ge V from the CDF measurement[26]
and 80.482 ± 0.091 GeV from the D0 measurement[27] lead to a combined
result of 80.448 ± 0.062. The CDF measurement uses both electrons and
muons, primarily in the central regions of the detector. The D0 measure-
ment is limited to electrons but exploits data from the end as well as the
central calorimeters. These results are in excellent agreement with the com-
bined results[28] of the four LEP experiments of 80.350 ± 0.056 Ge V. They
also demand the inclusion of electroweak loop corrections to be compat-
ible with the three basic electroweak parameters. The good behavior of
systematic errors discussed above is illustrated in Fig. 10 which contains
a projection of the evolution of the W mass uncertainty as the integrated
luminosity is increased even further with the upgraded detectors. One can
note that this projection preceded the most recent results but that these
Scaling of W-mass error

200 ~--------------,
Run lA. COF. Oil. UA2 (preliminary)
140
120
100
Run lb. CDF. 00 (anticipated)
80
60
- .. --'-'-'""'.'~
Scaling
40 "~< ,
+ reaolutJon
+ aya.ematlc. """',:<:«;..
20 .
4
JLdt (Pb'~O
Figure 10. Expected evolution of the precision of a measurement of the W-boson mass
at the Tevatron Collider.
100 pb- 1 measurements fit well on the curve. The cusp occurs as the num-
ber of interactions per crossing increases followed by a further change in the
time spacing between bunch crossings. It would seem that an uncertainty
of 40 MeV per experiment is not out of the question.
Because of its large mass, the top-quark is currently only directly acce-
sible at the Tevatron. The current data samples have allowed CDF and D0
to measure [29] the mass. The techniques vary depending on the channel
used. CDF uses final states in which the top and the antitop each decays
into three jets, two light quark jets from the intermediate W boson and a
b- or b- quark jet. However in this channel the background is large and the
resulting measurement has an uncertainty of around 10 GeV. In the dilep-
ton channel, even with two missing neutrinos, the mass can be determined;
however, the low statisistics offset a rather good understanding of the sys-
tematic uncertainties. Again from each experiment the uncertainties are of
order 10 GeV. For each experiment the dominant measurement comes from
the "lepton-pIus-jets" channel in which one of the intermediate W bosons
decays leptonically giving a charged lepton and a missing neutrino while the
other decays into light quarks. The final state then contains four jets and
a lepton for which each of the momentum vectors is fully measured, and
a neutrino for which only the transverse components are measured. Using
the mass constraints, those from the intermediate W bosons and that from
demanding that the top and anti top masses be the same, leads to kinematic
226 H.E. MONTGOMERY
.... LEP2 , pp Data

80.5 68% CL
:>
(J)
~ 80.4
3:
E
80.3
Preliminary
80.24-~~4-~~~~~~~~~~T-~
130 150 190 21 0
Figure 11. M w versus mt showing that a light Higgs is favoured by the current data.
fits with two constraints. Account has to be taken of the possible combi-
nations. These may be restricted if the b-quark jet is well identified, either
by a displaced decay vertex or from a soft charged lepton from the b-quark
decay. Using this channel, CDF achieves an uncertainty of about 7 GeV
including systematics whereas that from D0 is 8 GeV. All these measure-
ments are combined taking into account the correlations in uncertainties,
both those derived from common experimental errors (between channels in
a given experiment, and those derived from common techniques between
experiments. The result is that mt = 174.3 ± 3.2 ± 4.0 GeV. This makes
the mass of the top quark the best measured of all the quark masses.
We can then use the top mass and the W mass and compare them with
the basic electroweak predictions. As indicated earlier, the Higgs boson
mass enters into the calculations of the electroweak loops. Hence the com-
bination of all the electroweak measurements has sensitivity to the mass of
the putative Higgs boson. This is illustrated in Fig. 11; we see that the data
favour a light Higgs, of order 100-200 GeV. The uncertainties are rather
large but this tendency in the existing data both from the Tevatron and
from LEP jSLD has lent excitement to the searches current at LEP and to
work on the upgrades of the Tevatron experiments.
The top mass determination described above is dominated by the un-
certainties in the jet energy scale calibration. In future runs we expect that
these uncertainties can be reduced using the data themselves. CDF has ob-
served a W -mass peak in the decay jets from top and, in a bb data sample,
has observed the peak from the Z boson. Taking into account secondary
vertex triggers, which each experiment expects to use in the upcoming run-
ning, the latter will provide a powerful jet calibration tool. As a result
we can look forward to a reduction of the uncertainty on the mass of the
top-quark to less than 2 Ge V from each of CDF and D0 in the next few
years.
4.4. BEYOND THE STANDARD MODEL
It is the duty of experimenters at the highest energy colliders to search for

phenomena not previously observed. These searches are necessarily guided
by how we imagine the underlying physics. At any given time, certain sce-
narios enjoy more popularity than others. We have seen in the past that
states decaying into lepton pairs have often provided discoveries, for ex-
ample the J / 1jJ and the Z boson. QCD breaks the electroweak symmetry,
however not with sufficient strength to explain the masses of the Wand Z,
so some have postulated an analogous interaction, technicolor, to generate
such masses. At the present time, many theorists believe that supersymme-
try should playa role and the phenomenology of many possible scenarios
is well developed. As experimentalists, we keep an open mind.
Searches have been performed for the leptonic decays of higher mass
gauge bosons both W -like and Z-like. None have been found with masses
less than about 600-700 Ge V. Similarly, by looking in the mass spectra of jet
pairs, excited quarks can be excluded with masses less than about 600 GeV.
However, such states can also influence angular distributions of lepton or
quark pairs at energies well below their masses. Thus such measurements
provide windows to very high masses. A recent example is the measure-
ment[30] of lepton pair production which sets limits of 3-6 TeV on such
compositeness scales. A few years ago, an excess of high mass events was
observed[31] by the HERA experiments. One possible explanation was the
existence of 1st generation leptoquarks, composite electron-quark states.
Extensive measurements[32] at the Tevatron now exclude such particles
unless their masses exceed about 220 GeV. Similar, if slightly weaker, lim-
its have been place on the masses of 2nd and 3rd generation leptoquarks.
Technicolor searches are relative newcomers to the Tevatron analysis
menu. Searches have been performed[33] which provide lower limits on
228 H.E. MONTGOMERY
00 Preliminary, 1.07 Ib-'

>
'"
Cl1500 (a) Kinematic cuts
'"~1000
Q)
>
500
...
UJ
,.., ~
00 20 40 60 80 100 120 140 160
Dijet mass (GeV)
>.,
Cl 30 (b) Topological cuts
+ Single b-Iag
~ 20
.,
C
>
UJ 10
,-+-.
00 20 180 200
>Q) 60
Cl (c) Topologica I culs
<C
~
40
+t + Single b-tag
C
~
w
20 +
00 50 1 00 150 200 250 300 350 400
W + dijet mass (GeV)
Figure 12. Mass spectra expected for technicolor signals for the 'lrT, with a mass of 110
GeV, and the PT, with a mass of 210 GeV. The points are the expected distributions
including the technicolor signal; the shaded histogram is the standard model background.
the masses of technipions( 1fT ), technirho(PT) and techniomega( WT) mesons.

The techniques are illustrated by Fig. 12 in which the submass distribu-
tions of the cascade decay of a technirho( PT --+ 1fT + W --+ b + b + W) are
plotted. With the 1 fb- 1 used in this study[34] the signals at 110 and 210
Ge V are clearly visible.
TABLE 1. Mass ranges covered for a 50' discovery in SUSY models.

Model
I SUSY Particle I Mass
Run I( 0.1 fb- Run 11(2.0 fb- l
l
Limit(GeV) Mass Limit(GeV)
SUGRA I
-± 70 210
Xl
9 270 390
tl(-+ bX~) 170
GMSB
150 265
120
The most general Supersymmetry(SUSY) theories have more than one

hundred parameters and a comprehensive search is almost impossible. The
usual strategy is to search for signals suggested by particular classes of
models in which theoretical bias is applied to reduce the numbers of pa-
rameters to a few. Searches at the Tevatron initially concentrated on the
so-called minimal-Supergravity(m-SUGRA) models by looking for multi-jet
final states with missing transverse energy. The missing transverse energy
is supposed to be a clear signal for the Lightest Supersymmetric Parti-
cle(LSP); if R-parity is conserved, the LSP survives the decay chain, is
neutral, and will escape detection since its interactions with the matter
of the detector are weak. These have been complemented by searches in
channels containing leptons in the final state. The current limits[35] are in
the region of squark and gluino masses of 270 Ge V if the two are approxi-
mately equal and squark masses less than about 150 GeV are excluded for
any gluino mass.
If R-parity conservation is not assumed, missing transverse energy is no
longer a useful discriminant. Nevertheless searches have been performed[36,
37] particularly in channels with leptons in the final state. Limits on the
squark masses in such analyses are about 260-280 GeV, and similarly for
the gluino when the masses are equal.
In recent years alternative SUSY breaking scenarios to the SUGRA
models have been explored. In particular, the observation[38] of a single
spectacular e"e",,$T event prompted considerable activity. Theorists at-
tempted to accomodate the event using Gauge Mediated Symmetry Break-
ing(GMSB) which in general leads to numerous electromagnetic objects,
photons and electrons, in the final state along with missing transverse en-
ergy. Within the phase space of these models, limits on the mass of the
lightest chargino, of 120 GeV from CDF[38] and of 150 GeV from D0[39],
have been reported.
Looking to the future, a summary of what we can expect with about
2 fb- 1 is displayed in Table 1. These are the results of a fairly extensive
study[40] of the potential for future SUSY exploration at the Tevatron.
4.5. THE HIGGS
Almost independently of theoretical religion, it is believed that some spin

o high mass structure, known generically as the Higgs boson, must exist to
generate the masses of the Wand Z bosons. In its simplest form it could
have just one observable state. Analysis of the current electroweak data
suggest, as we have seen, that it be relatively light. If so it could potentially
be discovered in the current running of LEP 11[41], or, as concerns us here,
at future running of the Tevatron.
230 H .E. MONTGOMERY
I.'r---------,
I. ,
~
E<J9 ....... H
lJ(pp ~ H + x) (pbl
<I. - 2 reV
Ii " ; '
10~'---'---100'---'---'50'--...L----'200
... :
10 50 100
MHIGeVj
Figure 13. Higgs-boson production cross sections and branching fractions to fermions
and to bosons as a function of Higgs-boson mass.
The standard model Higgs particle may be produced at the Tevatron

in several different ways. As indicated in Fig. 13, the cross section for the
gluon fusion process is about 1 pb at 100 GeV. The "associated production"
of an electroweak gauge boson, W or Z, and a Higgs boson is approximately
an order of magnitude less. The decay branching ratios to fermions and to
bosons are also shown in the other two diagrams in Fig. 13. As expected,
at low masses the bb mode dominates. However one notices that the WW
mode becomes large for masses in excess of 130 GeV. Folding these facts
together and taking into account the need for a distinctive signature on
which to trigger, and on which to key the analysis, two distinct approaches
are discussed.
At low masses the approach is to look for a signal in associated pro-
duction exploiting both the decays of the W or Z and the bb or II decay
modes of the Higgs.
Current searches for a bosophilic Higgs use two jets[42]' or two lep-
tons[43] from the Wand Z decays, along with two photons from the Higgs
decay. D0 and CDF respectively place lower mass limits in the region of 80
GeV. Using the bb mode as a signal for the Higgs, and including the vIi de-
cay among the lepton decays of the Z, limits relevant to the standard-model
Higgs are obtained[44, 43]. These do not give a mass limit but constrain
the cross section at about ten times its standard model expected value.
This illustrates the premium on integrated luminosity. In SUSY models the
Higgs structure is more complicated and a charged Higgs exists. If its mass
would be less than that of the top quark then the H±b decay mode would
compete with that of the top quark. The extent to which this would occur
is controlled by the tan f3 SUSY parameter. The agreement between the
observed cross section times branching ratio for the top quark into modes
containing a W boson and the theoretical prediction has therefore been
used to place limits[45, 46] in the space of tanf3 and mH o
With the upgraded detectors, the efficiencies are improved by a factor
combined CDF /00 thresholds

95% CL limit
3a discovery
5a discovery
80 100 120 140 160 180 200

Higgs mass (GeV / c2 )
Figure 14. Luminosity required as a function of Higgs mass to achieve different levels
of sensitivity to the standard-model Higgs boson.From the upper curve corresponds to a
5 (J discovery, the middle a 3 (J signal and the lower a 95% exclusion limit. These limits
require two experiments, Bayesian statistics are used to combine the channels and include
the improved sensitivity which would come from multivariate analysis techniques.
of about four as a result of the improved detection of b quarks, and, per-

haps key, the ability of each experiment to trigger on displaced vertices.
Recent studies[40] take into account these improvements and improved un-
derstanding of the jet calibrations and resolutions. They find that in the
relatively low mass region, below about 130 GeV, the prospects for observa-
tion of a standard model Higgs are promising. If it turns out to be possible
to use all of the leptonic and haronic decay modes of Wand Z along with
just the bb decay mode for the Higgs, exclusion up to 140 GeV and a three
standard deviation hint up to about 130 GeV could be obtained with 10
fb-l. For higher masses, the key appears to be the use[47] of the WW* and
similar decay modes. Life is not easy when multiple modes are necessary
for the observation, nevertheless, the observation of the top quark was con-
siderably strengthened by such techniques. A 95% exclusion up to 180 Ge V
might be achievable with the same 10 fb- 1 of integrated luminosity. A hint
of a signal anywhere up to a mass of 180 GeV probably requires 20 fb- 1
of integrated luminosity. This is not excluded and is a challenge which the
experiments and the accelerator are keen to accept.
232 H.E. MONTGOMERY
5. Conclusions
In these lectures we have attempted to describe the Tevatron Collider com-

plex, the experiments CDF and D0 currently undergoing major upgrades,
the physics that has come out of 100 pb- 1 of integrated luminosity and to
give a sense of the prospects for the future. It is safe to say that much of the
physics at the Tevatron Collider has been a revelation. The experimental
environment has proved to be tractable. The events are busy but the ob-
jects relevant to physics, characterised by transverse momenta in excess of
100 GeV and of masses in the range 100 to 500 GeV, are readily observable
and measurable. In particular it has proved possible to calibrate the exper-
iments and the analyses so that precision measurements of the W mass,
the top quark mass and the jet cross section have been completed. Finally,
complete B states have been reconstructed, and their lifetimes measured
and a first measurement of the CP violation parameter sin 2(3 has been
made. If any doubts existed a decade ago as to the breadth of the potential
of the Tevatron, none should exist today.
6. Acknowledgements
In this electronic age we are used to having access at the push of a button
to descriptions and diagrams produced in many parts of the world. This
has greatly facilitated the preparation of these lectures. I would therefore
like to thank my many colleagues on the CDF and D0 experiments who
have helped in this work either knowingly or unknowingly. John Ellison,
Paul Grannis and Nick Hadley were kind enough to read the manuscript
and suggest corrections. Finally the school was immensely enjoyable and I
would like to express my appreciation to the organisers, the other speakers
and to the students from all of whom I learned much.
References
1. CDF Collaboration Home Page, http://www-cdlfnal.gov/
2. D0 Collaboration Home Page, http://www-dO.fnal.gov/
3. Montgomery, H.E. (1999) Recent Results from the Tevatron Collider, Yad. Phys.,
Vol 62, pp. 1-9, FERMILAB-Conf-99/056-E.
4. Montgomery, H.E. (1999) Physics at the Main Injector Proceedings, DPF99
U.C.L.A. ,FERMILAB-Conf-99/057, hep-ex/9904019.
5. Montgomery, H.E. (1998) The Physics of Jets, Ecole de Gif-sur- Yvette, Centre de
Physique des Particules, Marseille, France, FNAL-Conf-98/398.
6. Blazey G. & Flaugher B. (1999) Inclusive Jet and Dijet Production at the Tevatron
submitted to Ann. Rev. of Nucl. Part. Sci., hep-ex/9903058.
7. Abbott B. et al.,D0 Collaboration (1999) Inclusive Jet Cross Sections in pp Col-
lisions at VB = 630 GeV and 1800 GeV, paper submitted to the International
Europhysics Conference on High Energy Physics, EPS-HEP99, 15-21 July, 1999,
Tampere, Finland and to the XIX International Symposium on Lepton and Photon
Interactions at High Energies.

8. Asakawa T. for the CDF and D0 Collaborations (1999) Dijet Results from CDF
and D0 Proceedings of the XIII Topical Conference on Hadron Collider Physics,
Tata Institute of fundamental Research, Mumbai , India, 14-20 January 1999, World
Scientific, 1999.
9. Ellis S.D. and Soper D.E. (1993) Phys. Rev., D48, 3168.
10. Abbott B. et al., D0 Collaboration (1999) Subjet Multiplicity in Quark and Gluon
Jets at D0 , paper submitted to the International Europhysics Conference on High
Energy Physics, EPS-HEP99, 15-21 July, 1999, Tampere, Finland and to the XIX
International Symposium on Lepton and Photon Interactions at High Energies, hep-
ex/9907059.
11. Abbott B. et al., D0 Collaboration (1999) to be published in Phys. Rev., FERMILAB
PUB-99/171-E, hep-ex/9906025.
12. Affolder T. et al., The CDF Collaboration (1999) Submitted to Phys . Rev. Lett. ,
FERMILAB-PUB-99/220-E; Abe F . et al., The CDF Collaboration (1999) Phys.
Rev., D59, 052002.
13. Abbott B. et ai, D0 Collaboration (1999) Measurement of the Transverse Mo-
mentum Distribution of Wand Z Bosons Produced in pp Collisions at .;s = 1.8
TeV,paper submitted to the International Europhysics Conference on High Energy
Physics, EPS-HEP99, 15-21 July, 1999, Tampere, Finland and to the XIX Inter-
national Symposium on Lepton and Photon Interactions at High Energies, hep-
ex/9907044 .
14. Abe F . et al., The CDF Collaboration (1998) Phys. Rev. Lett., 81, 1367.
15. Nakada T. (1999) Lectures at this School
16. Buras A. (1996) TUM-HEP-255-96, Sep 1996, hep-ph/9609324.
17. Abbott B. et al., D0 Collaboration (1999) Forward Rapidity Muon and b-quark
Production in pp Collisions at .;s = 1.8 TeV, Paper #453 submitted to the XXIX
International Conference on High Energy Physics - ICHEP98, July 23-29, 1998,
Vancouver, B.C. , Canada.
18. Abe F. et al., CDF Collaboration (1998) Phys. Rev. Lett., 81, 5513.
19. Gershtein S.S., Kisilev V.V., Likhoded A.K., Tkabladze A.V. (1995) Uspeki Phys.
Nauk., 165, pp. 3-40., IHEP 94-81 , 1994.
20. Abe F . et al., The CDF Collaboration (1998) Phys. Rev. Lett, 82, 3576; Paulini M.
(1999) Int . J. Mod. Phys., A14, 2791.
21. Affolder T. et al., The CDF Collaboration (1999) submitted to Phys. Rev. D,
FERMILAB-PUB-99 /225- E.
22. Affolder T. et al., The CDF Collaboration (1999) Submitted to Phys. Rev. Lett.
FERMILAB-PUB-99/257-E.
23. Chakraborty Dhiman for the CDF & D0 Collaborations (1999) Study of the Top
Quark at the Tevatron , Proceedings of the XIII Topical Conference on Hadron Col-
lider Physics, Tata Institute of fundamental Research, Mumbai, India, 14-20 Jan-
uary 1999, World Scientific, 1999.
24. Ellison J. and Wudka J. (1998) Study of Trilinear Gauge Boson Couplings at the
Tevatron Collider Ann. Rev. Nucl. Part. Sci., 48, 33.
25. Rimondi F. CDF and D0 Collaborations (1999) W -t TV and rw Studies at the
Tevatron Collider Proceedings of the XIII Topical Conference on Hadron Collider
Physics, Tata Institute of fundamental Research, Mumbai , India, 14-20 January
1999, World Scientific, 1999.
26. Lancaster M. (1998) Proceedings of Results and Perspectives in Particle Physics, La
Thuile, Italy.
27. Abbott B. et al., D0 Collaboration (1999) to be published in Phys . Rev. Lett,
FERMILAB PUB-99/259-E, hep-ex/9909030 ; submitted to Phys. Rev. D., hep-
ex/9908057, FERMILAB-Pub-99/237-E) .
28. LEP Electroweak Working Group (1999)
http://www.cern.ch/LEPEWWG/plots/summer99/
234 H.E. MONTGOMERY
29. Sliwa K. (1999) Top Mass and Cross Section Results from CDF and D0 at the
Fermilab Tevatron Proceedings of the XIII Topical Conference on Hadron Collider
Physics, Tata Institute of fundamental Research, Mumbai, India, 14-20 January
30. Abbott B. et al., D0 Collaboration (1999) Phys. Rev. Lett., 82, 4769.
31. Cashmore R. (1999) Lectures at this School
32. Hagopian S. for the D0 and CDF Collaborations (1999) Leptoquark Summary
from the Tevatron Proceedings of the XIII Topical Conference on Hadron Collider
Physics, Tata Institute of fundamental Research, Mumbai, India, 14-20 January
33. Abe F. et al., The CDF Collaboration (1999) Phys. Rev. Lett., 83, 3124; Affolder T.
et al., The CDF Collaboration, (1999) submitted to Phys. Rev. Lett., FERMILAB-
PUB-99/141-E. ; Abe F.et al., The CDF Collaboration (1998) submitted to Phys.
Rev. Lett., FERMILAB-PUB-98/321-E.
34. Eichten E., Lane K., Womersley J., (1997) Phys. Lett., B405, 305.
35. Pagliarone C. (1999) SUSY Searches at Tevatron Collider Proceedings of the XIII
Topical Conference on Hadron Collider Physics, Tata Institute of fundamental Re-
search, Mumbai, India, 14-20 January 1999, World Scientific, 1999.
36. Abe F. et al., The CDF Collaboration (1999) Phys. Rev. Lett., 83, 2133; Abe F.
et al., The CDF Collaboration (1998) submitted to Phys. Rev. Lett., FERMILAB-
PUB-98/374-E.
37. Abbott B. et al., D0 Collaboration (1999) Search for R-parity Violating Super-
symmetry in the Dielectron Channel, to be published in Phys. Rev. Lett., hep-
ex/9907019.
38. Abe F . et al., The CDF Collaboration (1998) Phys. Rev. Lett., 81,1791; Abe F. et
al., The CDF Collaboration, (1999) Phys. Rev., D59, 092002.
39. Abbott B. et al., D0 Collaboration (1997) Phys. Rev. Lett., 78, 2070; Abbott B. et
al., D0 Collaboration (1999) Phys. Rev. Lett., 82, 29.
40. Physics at Run II - Supersymmetry/Higgs, http://fnth37.fnal.gov/susy.html
41. Fernandez H., (1999) Lectures at this School.
42. Abbott B. et al., 1999, D0 Collaboration (1999) Phys. Rev Lett., 82, 2244.
43. Wu X. for the CDF and D0 Collaborations (1999) Higgs Search at the Tevatron Pro-
ceedings of the XIII Topical Conference on Hadron Collider Physics, Tata Institute
of fundamental Research, Mumbai, India, 14-20 January 1999, World Scientific,
1999.
44. Abe F. et al., The CDF Collaboration (1998) Phys. Rev. Lett., 81, 5748.; Abe F.
et al., The CDF Collaboration (1998) Submitted to Phys. Rev. Lett., FERMILAB-
PUB-98/252-E.
45. Abe F. et al., The CDF Collaboration (1997) Phys. Rev Lett., 79, 357.
46. Abbott B. et al., D0 Collaboration (1999) Phys. Rev. Lett., 82, 4975.
47. Han T, Turcot A. and Zhang R. (1999) Phys.Rev. D59 093001.
HEAVY GAUGE BOSON PRODUCTION AT SMALL
TRANSVERSE MOMENTUM IN HADRON-HADRON
COLLISIONS
A. KULESZA
AND
W.J. STIRLING
Departments of Physics and Mathematical Sciences
Abstract. In this talk we briefly discuss the present status of the theoreti-
cal description of vector boson production at hadron colliders. In particular
we focus on resummation as a key method of improving the description
of Wand Z production at small transverse momentum. We present two
techniques for performing the resummation of large logarithms: impact pa-
rameter and transverse momentum space resummation. We also discuss the
benefits and shortcomings of both methods.
1. Introduction
The production of vector bosons at hadron colliders, i.e.
pp,pp --+ W±, Z +X , (1)
has long been considered an important process for our understanding of

QCD. Primarily a QCD test bed, it provided one of the first pieces of ev-
idence for the existence of gluons. Nowadays a comparison between data
and theory allows for a precise extraction of the W boson mass. Moreover,
the abundance of experimental data from the Tevatron necessitates the
development of a consistent theoretical description of the process. Such a
235
236 A. KULESZA AND W.J. STIRLING
description will be crucial for the analysis of data collected by the Large
Hadron Collider, where even larger event samples are expected. A precise
knowledge of Mw has profound implications for Higgs boson searches at
the Tevatron and LHC. One of the methods of obtaining Mw at hadron
colliders is to analyse the lepton transverse momentum distribution in W
production. A precise knowledge of the W transverse momentum distri-
bution is a crucial ingredient in this analysis. In this write-up we briefly
review currently available techniques for performing W transverse momen-
tum distribution calculations in QCD, and discuss their advantages and
disadvantages.
2. DLLA
The production of vector bosons gained an interest in the very early days
of QCD as a process to discriminate between the parton model and QCD
field theory. Clearly, within the framework of the parton model, a boson
produced in a qij collision would not have any transverse momentum qT
(we neglect here the intrinsic transverse momentum due to the Fermi mo-
tion). QCD predicts emission of gluons by quarks before the collision, which
manifests itself in a nontrivial distribution of qT. Fixed-order perturbative
calculations have been performed up to O(a;) (for a review of the litera-
ture see [5]), corresponding to 0, 1 or 2 gluon emission, and are in excellent
agreement with the data at large values of qT . The situation becomes dra-
matically different for small values of qT, where the fixed-order theoretical
predictions diverge rapidly as qT -+ O. The singular behaviour is easily vis-
ible already at the LO level and originates from one real gluon emission,
Fig la. The contribution to the differential cross section from these real
diagrams qij --+ V 9 is of the form
(2)
where A, B are calculable coefficients and C(q}) is an integrable function.

Thus the logarithmic structure of (2) is responsible for the divergence at
small qT .
To make matters worse, the emission of a larger number of gluons
(i.e. higher-order terms) causes the logarithmic divergence to become even
stronger. An example diagram representing a contribution coming from the
emission of only real gluons is shown in Fig. lb. It can be shown that in the
approximation of soft and collinear gluons with strongly-ordered momenta,
i.e.
(3)
HEAVY GAUGE BOSON PRODUCTION... 237
(.) (b)
Figure 1. a: Real diagrams for the LO process (2), b: One of the diagrams presenting real
gluon emission contributing to DLLA.
the dominant contribution yields [1]
(4)
with A(l) = 2Cp = 8/3. This approximation is commonly known as the

Double Leading Logarithm Approximation (DLLA) which is due to the
a~ln2N-l(Q2/q}) structure. l For a s ln2(Q2/q}) 1 the series in (4) di-
'"V
verges, i.e. the higher-order terms become dominant. In the case of the
W, Z production, Q M w , Mz, this corresponds to qT
'"V 10 -;- 15 GeV.
'"V
The bulk of the collected data lies below this limit. A solution to the diver-
gent behaviour problem is indicated by the structure of (4) itself. A closer
inspection shows that (4) has an exponential power series structure and
can be resummed, yielding a so-called Sudakov form factor [1]:
(5)
Resummation of the divergences in (4) to all orders results in a drastically

different behaviour of (5) as compared to the LO expression (2), cf. Fig. 2.
In particular, the resummed cross section vanishes in the small qT limit.
However, this suppression as qT --* 0 is unphysical, since it is a direct result
of imposing the strong ordering constraint (3) on the gluon transverse mo-
mentum phase space which, according to (3), vanishes in the qT --* 0 limit.
Clearly in this region one may expect (formally) sub-leading effects to be
important and lead to a non-zero cross section. In particular, the strong
lSometimes it is also called the first 'tower' of logarithms; 'i-th tower ' meaning an
infinite set of a;V In 2N - i (Q2 /q}) terms.
:;--
G
"0
'-
b
10 5
"0
2- 10 4 _ DlLA
;
:::- 10 3
_ ___ O{a.}
10 2
10
Figure 2. Comparison of LO perturbative calculation (only logarithmic part) and

DLLA (5), for the parton subprocess.
ordering condition does not take into account the overall transverse momen-
tum conservation, which allows, for example, a vector boson with small qT
to be produced in association with two almost back-to-back gluons with
non-negligible transverse momentum. In practice, the only requirement for
the production of a vector boson with qT '" 0 is that the vector sum over the
gluons' momenta, L:i qTi' is small. An obvious solution to the 'suppression
problem' is therefore obtained by imposing conservation of the transverse
momentum directly in the calculations. This is done with the help of the
b-space technique.
3. b space
The b-space method relies on performing a significant part of the calcula-
tion in impact parameter (b) space which is Fourier conjugated with respect
to the transverse momentum. First the cross-section (with a delta func-
tion conserving transverse momentum implemented) is transformed into
the Fourier conjugate, b-space,
~ (b) -
(7 -
jd 2qTeibqr (~~)
d 2 ' (6)
(To qT
then the conjugated cross-section a(b) is computed and finally transformed

back to momentum space
(7)
HEAVY GAUGE BOSON PRODUCTION ... 239
A great advantage of the b-space method is that the inclusion of overall

transverse momentum conservation is straightforward, via [2]
(8)
This method allows for a development of a general expression resum-

ming all terms of the perturbation series which are at leas~ as singular as
l/q'f when qT -+ 0. This reads [3] (for the parton level subprocess):
da = ao roo bdbJo(qTb)eS(b,Q2) , (9)

dq'f 2 io
where
(10)
(11)
47ra 2
bo = 2 exp( -,E) , ao = - - .
98
The first two coefficients in each of the series in (11) can be obtained [4]
from the exact LO+ NLO perturbative calculation in the high qT region by
comparing the logarithmic terms therein with the corresponding logarithms
generated by the first three terms of the expansion of exp(S(b, Q2)) in (9),
A(l) = 2Gp
A(2) 2Cp (N (~~ _ ~2) - 190TRnj)
-3Cp
~- + GpN (1917r2 - ~923 + 6({3))
c; -
G} (7r 2 - 12((3))
+ GpTRnj ~7r2) ,
°
with N = 3 and TR = 1/2. From Fig. 3 it can be seen that the b-space
result leads to a finite intercept at qT = for the distribution, i.e. it cures
the problem of small qT suppression of the DLLA.
°
Although the b-space method succeeds in recovering a finite , positive
result in the qT -+ limit, it suffers from two major drawbacks, related to
_OlLA
~ _ __ b-space
10
1 -
-1
10
-2
10
Figure 3. DLLA predictions (5) compared to b-space at the parton level. Here
TJ=q}/Q2.
the matching between perturbative and resummed results and dealing with
the non-perturbative regime of very small qT. The resummed expression for
W, Z production is supposed to describe the cross section well in the limit
of small qT, typically 17-2 ~ qT ~ 10 7- 15 GeV.
For higher values of qT the logarithmic parts of the terms in the per-
turbation series cease to dominate the remaining, 'finite', parts. Also, the
value of the coupling decreases so that higher-order terms in as become less
relevant. Therefore, in order to obtain a consistent description of data one
needs to match these two predictions [6] at some intermediate value of qT.
By definition, the fixed-order and resummed calculations take into account
different sets of terms of the perturbation series. To perform matching it is
necessary to identify these sets, since then, by means of simple subtraction,
one can avoid double-counting. In the case of the b-space resummation
one automatically resums all known logarithmic terms (i.e. containing log-
arithms of In( Q2 / q})) and there is no way to select any particular subset
of them. Hence there is no unambiguous prescription for matching; existing
prescriptions require 'unsmooth' switching from resummed to fixed-order
calculation at some value of qT. This results in unphysical predictions for
the cross section around the matching point.
Smaller values of qT (i.e. smaller than 1 GeV) belong to the non-
perturbative regime and as such their treatment poses an unsolved the-
oretical problem which is normally overcome by introducing some kind
of parametrization of the non-perturbative effects in the cross section ex-
pression. In particular, the integration in (9) extends from 0 to 00. Thus
when b -+ 1/ A, as(l/b) becomes large and one enters the non-perturbative
regime. In other words it is impossible to make predictions for any qT with-

out having a prescription for how to deal with the non-perturbative regime
of large b. One approach is to artificially forbid b from reaching large values
by replacing it with a new variable b* which serves to 'freeze' the pertur-
bative calculations at a certain point bJim,
b
b* = -Jr==1=+:::;::(b;:=;:/b:=li=m~)2
This procedure allows the cross section to be computed but the resulting
prediction is not in agreement with the data at very small qT. This can be
achieved by multiplying the integrand by an arbitrary (non-perturbative)
function F:t P (Q, b, x A, x B), such that the resulting cross section
can, with a particular choice for the function F:t P, describe the data rea-
sonably well for very small qT. It is clear from (12) that F:tP(Q,b,XA,XB)
affects the cross section for all values of qT, a feature that one would like to
avoid. More importantly, the detailed form of the non-perturbative func-
tion F:t P (Q, b, x A, XB) remains a matter of theoretical dispute. However,
the (fixed-target Drell-Yan) data suggest that the form of pNP is approx-
imately gaussian in qT-space, which implies also a gaussian form for the
non-perturbative function in b-space. For a discussion of the form of F:t P
the reader is referred to [5].
4. qT-space method
A question which naturally arises out of the above discussion is whether it is
possible to reproduce all the good features of b-space resummation without
the drawbacks related to this method. It turns out that one can indeed
achieve this by performing the calculations directly in qT-space. First, in
qT-space it is straightforward to distinguish between terms which are or
are not resummed. Therefore it should be more straightforward to perform
matching with the fixed-order calculation, and the final result should be
smooth near the matching point. Moreover, the non-perturbative input
would be required in, and would affect only, the small qT region.
There has recently been a great deal of theoretical interest in developing
a resummation technique in qT-space, and as a result three methods have
been proposed [7, 8, 9J . All of them resum different subsets of logarithmic

terms (for a discussion see [10]) but a common starting point is the b-space
expression (9). Here we discuss in detail only one such approach (KS),
as derived in [9], which allows us to fully resum the first four 'towers' of
logarithms.
In the simplest case, when the coupling as is fixed and only the leading
coefficient A(l), i.e. A(as) = asCF/,rr, B(as) = 0, is retained, the final
result in this KS approach reads
1 da >. -).. L2
- - = -e 2 x
ao dry ry
"fl
00 (_2>.)(N-l) N-l (
(N - I)! ~O
N - 1)
m
[
L N - 1 - m 2TN+m + LTN+m-l
]
. (13)
Here L = In(l/ry) = In(Q2/q}), >. = asCF/,rr, ry = q}/Q2, and the numbers

Tm are defined by
rOO y
Tm == 10 dyJl (y) lnm( bo) (14)
The Tm can be calculated explicitly using the generating function
L ~m
00 _ tI'Er(I+~) _
(_ i) - exp
[_ ((2k+l) (!)2k+
2 L 2k 1 2
l] '
00
m=O m .
,t Tm - e
r 1 2 k=l + (15)
so that e.g. TO = 1, Tl = T2 = 0, T3 = -~((3) etc. Naturally, for numer-

ical calculations based on the expression (13) it is necessary to introduce
a cut-off value N max on the number of terms retained in the 'all-orders'
perturbation series. This determines the number of fully resummed 'tow-
ers' of logarithms to be also N max . However, it holds only in the case of
all coefficients in the A, B series, except A(l), being absent. In reality, due
to the lack of knowledge of A (3), B(3), etc. it is only possible to obtain a
result where not more than the first four 'towers' of logarithms are fully
resummed.
Equation (13) is clearly different from the DLLA expression (5). How-
ever the DLLA result can be recovered from (13) by setting all Tm coeffi-
cients (except TO) to zero. In fact, the Tm logarithms with m > 0 correspond
to taking into account the transverse momentum conservation which is au-
tomatically 'built in' to the b-space approach. The larger is N max , the more
of these 'kinematic' logarithms are taken into account, and the better the
b-space result is approximated. This is illustrated in Figure 4, which shows
how for small values of qT the approximation of the b-space result improves
with an increasing N max . Since in this calculation only A(l) is kept non-
zero and the coupling is fixed, there are no sources of subleading logarithms
.,.
"0
"-
b
"0 10 3
,;
:::- 10 2
10
- b sp oc e
• •. . N_.=l ( Sudokov )
10- 1 N_. = 5
-3
1010L_~8~~7~~6~~5~=L-~4~10L_~3~1~0-~2~1~0--1~~
10- 10- 10- 10 1
1)
Figure 4. qT-space predictions (13) for various values of N max together with the b-space
result (9). The DLLA result is recovered for N max = 1. Here 'T/ = q}/Q2.
in (9) other than those related to kinematics, i.e. related to conservation

of transverse momentum, preserved in the original b-space expression. By
explicitly including the T m coefficients in qT-space, therefore, we are able
to take transverse momentum conservation into account.
The technique developed so far enables us to derive a r' ealistic ' expres-
sion, which together with a proper treatment of kinematics includes also
sub leading logarithms corresponding to the running of the coupling and the
presence of the other known coefficients A,B (for further discussion see [9]).
After convoluting this parton level cross section with the parton structure
functions one finally obtains an expression for the gauge boson production
in hadron-hadron collisions. A comparison between the Tevatron DO data
[11] and two of the qT-space results (KS and EV [7]) , for the particular case
of Z production, is shown in Fig. 5. The difference in the two approaches
is that more subleading logarithms are taken into account in the KS cal-
culation. No matching or non-perturbative input has been assumed in the
theoretical predictions, but it is obvious from Fig. 5 that their presence is
necessary in order to improve the description of the data. It is certainly
possible to introduce a gaussian non-perturbative contribution to improve
the agreement .
5. Conclusions
We have briefly presented current issues concerning the theoretical descrip-

tion of vector boson production at small qT . It has been shown that resum-
700
:>
Q)
"
'-.
.0 600 DO Z data , absolute norma lizat ion
5 EV
2a: 500 KS
'-.
"
"0 400
300
200
100
a
-100
a 5 10 15 20 25 30 35 40 45 50
p,
Figure 5. Comparison of the (resummed part of the) theoretical predictions derived in

the Ellis-Veseli [7] and Kulesza-Stirling [9] approaches with DO data [11].
mation of large logarithms In( Q2 / q}) is necessary in order to obtain reliable

predictions in the low qT region. We have also shown that the transverse
momentum space technique for performing this resummation can overcome
some of the problems faced by the impact parameter method. However,
more work is required on this approach and on the qT-space method in gen-
eral. In particular, the form of the non-perturbative function in qT-space
still remains an open theoretical issue. However we believe that these prob-
lems will be eventually overcome and qT-space will provide a phenomeno-
logically successful description of vector boson production, i.e. a description
which will sufficiently well approximate the b-space result in the small qT
limit and match smoothly onto the fixed-order result in the large qT regime.
References
1. Dokshitzer, Yu. L. , Dyakonov, D. I. and Thoyan, S. I. (1980) Phys. Rep., 58, 269
2. Parisi, G. and Petronzio, R (1979) Nucl. Phys., B154, 427
3. Collins, J. Soper, D. and Sterman, G. (1985) Nucl. Phys., B250, 199
Collins, J. and Soper, D. (1981) Nucl. Phys., B193, 381; Erratum (1983) Nucl.
Phys., B213, 545
Collins, J. and Soper, D. 1982 Nucl. Phys., B197, 446
4. Davies, C. and Stirling, W. J. (1984) Nucl. Phys., B244, 337
5. Ellis, R K., Ross, D. A. and Veseli, S. (1997) Nucl. Phys., B503, 309
6. Arnold, P. B. and Kauffman, RP. (1991) Nucl. Phys., B349, 381
7. Ellis, R K. and Veseli, S. (1998) Nucl. Phys., B511 , 649
8. Frixione, S., Nason , P. and Ridolfi, G. (1999) Nucl. Phys ., B542, 311
9. Kulesza, A. and Stirling, W . J. (1999) Nucl. Phys., B555, 279
10. Kulesza, A. and Stirling, W. J. (1999) hep-ph/9909271
11. DO collaboration (1999) hep-ex/9909020, submitted to Phys. Rev. Lett.
RECENT DEVELOPMENTS IN THE THEORY
OF STRONG INTERACTIONS:
POWER CORRECTIONS AND RENORMALONS
G.P. KORCHEMSKY
Laboratoire de Physique Theorique t
Universite de Paris XI
F-91405 Orsay Cedex, France
1. Introduction
The goal of these lectures is to provide an elementary introduction into

a recent progress in understanding of nonperturbative hadronization phe-
nomenon in Quantum Chromodynamics.
It is widely believed that QCD is a true fundamental theory of strong
interactions. Being a part of the Standard Model, it is free of any ambigui-
ties or inconsistencies (like Landau pole problem in QED and Higgs search
in weak interactions). It has been tested with a unprecedented accuracy
over a wide interval of energies and up to now there are no indications that
at presently accessible energies QCD needs a revision. This does not mean
however that QCD has transformed into an "applied" theory. Even though
we know that the theory is true we do not understand in full how does it
work. One of the main puzzles of QCD has to do with hadronization phe-
nomenon and it can be formulated as follows. It is well known that on the
theory side the elementary quanta of strong interactions are colored quarks
and gluons. Their interaction is described by the SU(3) gauge theory
with as = g; /
(47f) being a strong coupling constant, Nc = 3 is the number
r
of quark colors, F:JL = oJLA~ - ovA~ + bc AtA~ a gluon strength tensor,
DJL = oJL - iA~ta and q! is the quark field of a flavor f = {u, d, s, c, b, t}.
tUnite Mixte de Recherche du CNRS (UMR 8627)
245
J.-J. Aubert et at. (eds.), Particle Physics: Ideas and Recent Developments, 245-266.
246 G.P. KORCHEMSKY
At the same time, on the experimental side the observed physical states in
QeD are not colored quarks and gluons but color neutral hadrons (mesons
and baryons). The bridge between two different "worlds" relies on the re-
markable properties of confinement and asymptotic freedom in QeD. The
interaction between quarks and gluons becomes weak at short distances,
or equivalently at large energy Q, while at large distances quarks and glu-
ons are confined into bound hadronic states. The effective QeD coupling
constant depends on the energy scale Q and it is given to one-loop approx-
imation by well known expression
(1.2)
with 130 = (131 Nc - ~n f) / (47f) being lowest-order approximation to the QeD

f3-function and nf the number of light flavors. The dimensionfull param-
eter A QCD is a fundamental QeD scale and its calculation out of the first
principles is one of the challenges in the theory of strong interactions. Al-
though one can not apply (1.2) to calculate the absolute value of O!s(Q)
without knowing AQCD, the evolution of the coupling constant with the
energy scale Q is under control. The numerical value of AQCD can be ex-
tracted from experimental measurements of different QeD observables and
it is given by
2
A QCD
1) ~ (0.2 GeV)
= Q 2 exp ( - f300!s(Q2) 2
. (1.3)
This expression implies that, firstly, AQCD is of nonperturbative origin and,

secondly, at high energies Q2 » A~CD one enters into a perturbative regime
of QeD, in which coupling constant becomes small O!s(Q) « 1 and certain
QeD observables (the so-called infrared safe quantities) can be approxi-
mated by perturbative series in O!s(Q2). This does not mean however that
including higher order perturbative corrections one can increase infinitely
the accuracy of QeD calculations. The reason for this is that there always
exist nonperturbative corrections due to hadronization phenomenon that
we do not know how to calculate yet. The success of perturbative QeD is
closely related to the fact that the latter corrections appear suppressed by
powers of the energy scale rv (AQCD/Q)P and at higher energies they are
expected to be small.
2. Power corrections in e+ e- annihilation
In what follows we shall concentrate on a particular example of QeD hard

process: e+e- annihilation into hadrons. This process has been measured
THE THEORY OF STRONG INTERACTIONS 247
1,Z
Figure 1. QeD evolution of the final state in e+ e- annihilation into hadrons
over a wide interval of the center-of-mass energy 8 and it served for many
years as a laboratory for testing QeD.
Analyzing the space-time picture of e+e- -t hadrons at large energy
8 one distinguishes three different regimes (see Fig. 1). First, electron and
positron annihilate to produce a vector boson (,* or Z) which in turn
decays into a pair of energetic quarks. Then, quark and antiquark move
back-to-back in the center-of-mass frame and initiate QeD branching at
short distances rv 1/8 through the subprocesses q -t qg, 9 -t gg, ... , which
one calculates perturbatively in powers of O:s (8). Perturbative QeD allows
us to describe the evolution of quarks and gluons up to the time scales at
which their energies approaches AQCD. At this moment one enters into a
nonperturbative regime, in which produced quarks and gluons are converted
into hadronic states. This transition occurs locally in the phase space -
hadrons are formed out of clusters of quarks and gluons having a small
invariant mass and propagating in the same direction. Finally, observing
produced hadrons one measures the (normalized) total cross-section
(2.1)
The dependence of Re+e- (8) on the energy 8 = E;'m is shown in Fig. 2

for 1 Ge V < Ecm < 60 Ge V. We observe that at small Ecm < 5 Ge V the
dominant contribution comes from a few lower mass hadronic states while
at higher energies the number of produced hadrons grows very fast leading
to a smooth dependence on the energy. To estimate the effect of hadroniza-
tion corrections one has to compare (2.1) with similar quantity (8)R;le-
calculated perturbatively on partonic level with quarks and gluons treated
248 G.P. KORCHEMSKY
as asymptotic physical states

Re+e-(s) = R~le-(s) + hadronization corrections (2.2)
with "state-of-the-art" perturbative result
R;:.';- (8) = N, ~ e} [1+ ",;8) +14 (",;8)) 2_12.8 (",;8)) 3+ O(,,!)]

(2.3)
One observes using Fig. 2 that the difference between parton (quark and
gluon) and hadron spectra is not important at high energy Ecm > 25 GeV
while it becomes significant at small s. This property is known as quark-
hadron duality and it expresses the fact that due to confinement phe-
nomenon the probability for quarks to become hadrons at high energies
equals 1 up to hadronization corrections which decrease with the energy as
a power
Hadronization corrections = Power corrections = O(A~CD/ sP) (2.4)
with p = 2. At the same time, hadronization corrections dominate at small
energy - in contrast with continuous perturbative spectra of quarks and
gluons, the physical spectra of hadrons is discrete.
Till recently the conventional QCD approach to calculating physical
observables like Re+e- (s) looked like
- Calculate as many terms of perturbative expansion as you can;
- Estimate hadronization corrections by turning on Monte-Carlo event
generators;
- Combine both and confront your QCD predictions with the experi-
mental data.
This procedure worked reasonably well until the following problems have
been identified. It was observed that perturbative QCD corrections grow
like n! to higher orders in as and therefore perturbation series in QCD
become divergent. In addition, Monte-Carlo event generators are "black-
boxes" with a large number of adjustable parameters. Turning them on one
arrives at the correction factors without gaining much in understanding the
hadronization effects. Finally, the "strength" of power corrections p depends
on the observable under consideration and for certain quantities like event
shapes in e+e- -annihilation (thrust, heavy mass, etc.) it takes the value
p = 1/2. A simple calculation shows that in this case hadronization correc-
tions become compatible with the second-order perturbative corrections at
high energy
a~(s) '" 1 ~V '" 0.01 for Vs = Mz = 91.2 GeV (2.5)

I{1.2
MAliK I
+- MAliK IILCIII
• I MilA
OAMY OCRYBTAL RALL DIAD!!: <>-MARK I I TOPAZ

~
.CKLLO ·;: :CUSS • LIINA ;,-PLUTO XVEIWS
T(nS)
II-Its.f
.CLEO .!.DAIlPll *MAC I{TASSO
1() 16 20 2.'i 30 311 40 45 50 II.'i 60
jI;.m (a.V)
Figure 2. Dependence of Re+ e- (s) on the center-of-mass energy. Solid lines correspond
to perturbative QeD fits.
and therefore should be systematically taken into account. Thus, in order to

improve the accuracy of QeD predictions one has to answer the following
three questions
- What shall we do with n! divergent perturbative QeD series?
- Do we have theoretical methods for estimating power corrections?
- How to combine consistently together perturbative QeD predictions
with nonperturbative power corrections?
It turned out that three problems become intrinsically interrelated. Namely,
as we will see, n!-divergences of perturbative QeD series follows some
universal pattern which one associates with the contribution of so-called
IR renormalon [1]. One should mention from the very beginning that IR
renormalon is not a physical state in QeD but rather an auxiliary object to
which one ascribes divergences of perturbation theory of a very particular
form
apT(S) = L a~(s)an, (2.6)
n
with p and a being some numbers and /30 defined in (1.2). We will show
that its appearance in perturbative series serves as an indication for the
250 G.P. KORCHEMSKY
leading power corrections to be of the form
(2.7)
with A 2p "" A~CD being some (nonperturbative) scale and the strength
of power corrections, p, governing asymptotics of perturbative coefficients
(2.6). Thus, the main idea behind IR renormalons is that the study of large
order behaviour of perturbative series allows to predict the structure of
the leading power corrections (critical exponents p) but not their absolute
magnitude (the scales Ap).
One should notice that it is not for the first time that one encounters
n!-divergencies in perturbation theory and discusses their relation to non-
perturbative effects. Similar phenomenon is well known both in the quan-
tum mechanics and field theory [2]. As we will see in a moment, instanton
contribution is one of the sources of n! divergences.
3. Perturbation theory at large orders: from Quantum Mechanics

to Field theory
Let us start discussing n! divergences in Quantum Mechanics [2]. The best
known example comes from the calculation of the ground state energy of
the double-well potential defined by I-dimensional Hamiltonian
(3.1)
with 9 being a coupling constant. One finds the ground state energy Eo{g)
by calculating the trace of the evolution operator and examining the limit
of large evolution time
tr e- TH T-::.,oo e-TEo(g) . (3.2)
Let us start with "naive" perturbative calculation by expanding Eo (g) in

powers of g2 around one of degenerate minima of the potential. In this way
one arrives at the perturbative series
(3.3)
where Eo(g = 0) term corresponds to the ground state energy of harmonic

oscillator. Examining large order asymptotics of perturbative coefficients
one discovers familiar n! behaviour
(3.4)
Its origin can be easily understood if one expands (3 .2) over the set of
Feynman diagrams and treats g-dependent terms as cubic and quadric
interaction vertices. Then, each individual Feynman diagram provides'" nO
contribution but the total number of diagrams grows as n! leading to (3.4).
Eqs. (3.3) and (3.4) imply that perturbative expansion for a well defined
physical quantity Eo(g) is divergent . The reason for this is quite obvious.
Calculating Eo(g) perturbatively we missed an important nonperturbative
phenomenon - tunneling under barrier. It is well known that the latter
effect which one usually interprets as a nonperturbative contribution of
quantum mechanical instantons (solution to the classical equation of motion
in imaginary time) produces an important correction to the energy 1
8Einstanton '" exp ( __1_) (3.5)

o 3g 2 '
which has an essential singularity at g2 = 0 and therefore cannot be calcu-

lated perturbatively. Nevertheless, one could wonder whether n! divergences
of perturbative series should serve as an indication of existence important
nonperturbative corrections. To examine this possibility one has to give
a meaning to the divergent series (3 .3). This can be achieved by noticing
that (3.3) is an asymptotic series in g2. This identification means that,
firstly, truncating an infinite sum and including a sufficiently large number
of terms, S N = L:;:;
E n g2n , one can use the perturbative expression S N to
approximate asymptotically Eo(g) but never calculate it exactly. Secondly,
the accuracy of the approximation is controlled by the first omitted term
(3.6)
Examining the behaviour of the r.h.s. one finds that thanks to small value
of the coupling constant EN9 2N decreases with N until some N = No and
then rapidly increases to infinity for N > No due to factorial growth of per-
turbative coefficients. Therefore, in order to get the best approximation to
Eo (g) one should not take into account too many terms but rather truncate
the series at N = No corresponding to the slowest growth of the partial sum
SN, or equivalently oN(ENg2N)IN=No = 0 (see Fig. 3). It is easy to see that
for EN having the asymptotic behaviour (3.4) the value of No is inversely
proportional to the coupling constant, No = 1/(3g 2 ). Substituting N = No
back into (3.6) one finds that one can not calculate Eo (g) perturbatively
1 More precisely this correction is produced by a pair of instanton-antiinstanton

252 G.P. KORCHEMSKY
N!
E
PT ~-~=-""'--'----------
best approx,
Figure 3. Dependence of the partial sum of Borel nonsummable asymptotic series

SN = L::=o Eng 2n on the number of included terms N.
with the accuracy better than
Eo(g) = EpT(g, No) + fl.Eo(g) ,

fl.Eo(g) '" 92N0 ENo '" exp ( - 3g 2 1)
(3.7)
where EpT(g, No) = L:n<No ENg 2N . Notice that the expression for fl.Eo(g)
matches exactly nonperturbative correction to the energy due to the in-
stantons, Eq. (3.5). Thus, analysing the perturbative series for Eo(g) we
encountered ambiguity that we can identify as an indication for the pres-
ence of nonperturbative corrections. The same property can be formu-
lated in the following way: perturbation theory generates nonperturbative
'" exp( -1/(3g 2 )) terms but fails to predict them uniquely. It is clear that
in order to calculate these terms exactly one has to apply nonperturbative
methods.
As we will see in a moment, the relation between n! divergences of
perturbative series and nonperturbative corrections is of general nature and
it holds also in QCD. The important difference with Quantum Mechanics
is that we do not have nonperturbative methods in QCD that could allow
to put nonperturbative effects under a control.
Analysing ambiguities of perturbative series it becomes convenient to
apply the so-called Borel transformation. This transformation amounts to
replacing the original series in a coupling constant by a new series in an
auxiliary (Borel) parameter eJ
=> B[E](eJ) = LeJnEn~l . (3.8)

n n.
Thus defined series has better convergency properties and having deter-
mined B[E](a) we could calculate E(g) through the integration over the
Borel parameter as
E(g) = 10roo dae- u / g 2

B[E](a). (3.9)
Using the large order behavior of the perturbative coefficients of a gen-

eral form Ek '" k!kab k with a and b being some parameters it becomes
straightforward to sum up the series (3.8) to arrive at
(3.1O)
Thus, n! divergences have transformed under the Borel transformation into

singularities of B[E](a) at a = lib. Depending on a sign of b, or equivalently
on the position of the Borel singularity on the real axis with respect to the
origin one has to distinguish two cases: b> 0 and b < O.
For b < 0 the perturbative coefficients grow at large orders as '" (-1 }nn!
and the series becomes sign alternating. As a consequence, the Borel singu-
larity is located outside the integration range in (3.9) and its existence does
not prevent us from performing integration in (3.9). Thus, for b < 0 the
series becomes Borel summable and one can apply different technique (con-
formal mapping, Pade approximation, Euler transformation) to improve an
accuracy of perturbative prediction for E(g).
In contrast, for b > 0 the series has a fixed sign factorial divergence'" n!
and the Borel singularity is located on the integration axis at a = lib. For
instance, for the double-well potential the singularity is situated at b = 3.
In this case, the Borel integral {3.9} is not well defined and in order to
give a meaning to E(g} one has to provide a prescription for integrating
the Borel singularity. Different prescriptions can be tried (principal value,
deformation of the integration contour, etc.), but their usage leads to the
results which differ from each other by the contribution proportional to the
residue at a = lib
(3.11)
with numerical prefactor depending on the prescription. There is no sur-

prise that going through the Borel transformation we recovered the same
perturbative ambiguity as in (3.5). One should notice that among all pos-
sible prescriptions there is only one that leads to a physically correct (and
therefore unique) result. Unfortunately, in order to define this prescription
one has to gain a full control under nonperturbative effects. This becomes
254 G.P. KORCHEMSKY
possible in Quantum Mechanics but in QeD the problem is still open. Nev-
ertheless, analysing ambiguities of perturbative QeD series we can obtain
some information about the possible structure of yet unknown nonpertur-
bative corrections but we are not able to make any quantitative predictions.
Turning from Quantum Mechanics to QeD we shall follow the same
route: identify n! divergences of perturbative series in QeD coupling con-
stant as, study the properties of resulting Borel singularities a = 1jb and,
finally, identify the ambiguities of the Borel integral as an indication of
existence of nonperturbative corrections of the form rv exp( -ljba s ). It
worth while to notice that we have already encountered such corrections
in QeD - the fundamental QeD scale (1.3) is exactly of this form. More-
over, taking into account that QeD coupling constant runs with the energy
scale as = a s (Q2) one realizes that nonperturbative corrections that we are
looking for are power corrections of the form
(3.12)
The "strength" of the power correction is determined by the position of

the Borel singularities and therefore can be extracted from the analysis
of n! divergences of perturbative series. At the same time, this correction
will enter into physical observable with a definite numerical prefactor that
remains still unknown and has to be calculated nonperturbatively.
Let us examine what are the possible sources of n! -divergences in per-
turbative QeD. One of the lessons that we have learned from the double-
well potential model is that classical solutions to the (Euclidean time) equa-
tions of motion generate Borel singularities in perturbative series. Similar
classical solutions exist also in QeD - the well-known instantons. Let us
examine their contribution to the partition function of QeD
(3 .13)
which is an analog of (3 .2) in Quantum Mechanics. Here for the sake of

simplicity we neglected quark contribution and didn't write explicitly gauge
fixing terms. It is easy to verify using the definition (3.9) , that the Borel
transform of the partition function is given by
(3.14)
To calculate the contribution of the classical solutions to the Yang-Mills

equation of motions A*(x) we split the gauge field AJL(x) = A~(x) + aJL(x)
and integrate over fluctuations a(x}. Taking into account that 6ASYM[A] =
o forA = A* we obtain
B[ZQCD](a} = J Va 6 (a - SYM[A*]- ~6~SYM[A*]a2) rv a_ S~M[A*] .

(3.15)
Thus, classical Yang-Mills configurations of the gauge field produce Borel
singularities at a equal to the value of the Yang-Mills action. A detailed
analysis shows that for A * (x) given by superposition of (multi}instanton-
antiinstanton pairs with topological charge zero, the action takes the dis-
crete set of values SYM[A*] = 47fn with positive integer n counting a number
of instanton-antiinstanton pairs. We conclude that QeD instantons gener-
ate the sequence of the Borel singularities located at
alnstanton = 47f, 87f, ... (3.16)
Similar to (3.10) and (3.11), each of them produces a power correction with
the leading contribution coming from the Borel singularity closest to the
origin. However, estimating this contribution at the energy scale given by
Z -mass one finds that it is negligibly small
A2 ) 47r!3o
rv
( ~~D rv 10- 512 (3.17)
for Q = Mz = 91 GeV . Thus, if QeD instantons were the only source of

nonperturbative corrections we should have expected their contribution to
be important only at lower energies. This fact contradicts to the data which
indicate an existence of sizeable power corrections even at high energies
Q = M z. We conclude that the origin of the latter corrections is not related
to QeD instantons and, as a consequence, we should look for another source
of n! divergences in QeD.
4. Renormalons
Let us consider the normalized total cross-section of e+e- annihilation into
hadrons defined in (2.1). According to the optical theorem it can be cal-
culated as imaginary part of photon polarization operator in QeD (see
Fig. 4)
R,+ ,- (8) ~ 12n ( ~>J) ImII(8 + iO), (4.1)
which in turn is defined through an expectation value of the time-ordered

product of quark electromagnetic currents JJ-I(x} = L.f~1 efiiJ (x}rJ-lqf (x)
(qJ-lqv - gJ-lvq2)II(l} = -i J
d4 xe- iqX (0IT {JJ-I(x)Jv(O)} 10). (4.2)
256 G.P. KORCHEMSKY
Polarization operator II(q2) can be calculated in perturbative QeD for

large negative q2 = _Q2 . For these values of photon virtuality the QeD
interaction occurs at short distances "-' I/Q and perturbative expansion in
powers of as (Q2) becomes meaningful. Then, to arrive at perturbative QeD
prediction for (4.1) one has to analytically continue perturbative expression
for polarization operator to positive q2 = S.
Figure 4. Relation between the total cross-section of e+ e- annihilation and imaginary

part of polarization operator
To the lowest order of perturbative expansion the polarization opera-

tor II(Q2) is given by quark loop that produces a conventional ultraviolet
divergence. This divergence does not contribute however to the imaginary
part of II( Q2) and it becomes convenient to introduce instead a UV finite
Adler D-function
(4.3)
Its perturbative expansion produces (an infinite) set of Feynman diagrams

each contributing with certain Feynman integral accompanied by color fac-
tor and additional power of nF counting the total number of quark loops.
As a consequence, the perturbative expression for D( Q2) has the following
general structure
= 1 + L a~+1(Q2)
00
D(Q2) [niDk,k + nt- Dk,k-l + ... + n~Dk,o]

1 , (4.4)
k=O
where the coefficients Dk,m = Dk,m (Q2, N c ) get contribution from the dia-
grams to the k-th order in a s (Q2) having m internal quark loops. Our goal
is to study the large order behavior of the perturbative coefficients Dk,m
as k -+ 00. It is clear however that since the total number of coefficients
Dk,O, Dk,l, ... , Dk,k grows fast with k, one has to find an approximation
that could allow to select a dominant contribution.
+ 2
Figure 5. Single renormalon chain diagrams
One of the possible approximations corresponds to the limit of large

number of quark flavors , nF » 1 and O:s(Q2)nF = fixed. Applying this
limit to (4.4), one finds that to the k-th order in O:s among k + 1 different
terms only one term survives '" Dk ,k containing the maximal number of
quark loops. It is easy to see that the corresponding Feynman diagrams
have a peculiar form shown in Fig. 5 to which one usually refers as bub-
ble diagrams or single renormalon chain. In these diagrams a single gluon
propagator carrying the momentum [Il- is "dressed" by N quark loops. The
contribution of each fermionic loop is given by well known QED like ex-
pression'" (30,FO:s(Q2) In ~ with (30 ,F = -inF coinciding with quark con-
tribution to the QeD (3-function, Eq. (1.2). Then, the expression for the
D-function in the large nF-limit looks like [3J
(4.5)
where integration goes over virtuality of gluon with Euclidean momentum

[Il- and the characteristic function F([2) describes the coupling of gluon to
the quark loop attached to external photons.
Examining the asymptotics of the integrand (4.5) at large n we find
that the dominant contribution comes from two regions: [2 « Q2 and [2 »
Q2, in which In(Q2j[2) becomes large. The behavior of the characteristic
function in these two regions is given by F([2) = 3~F ~ for [2 « Q2 and
F([2) = -¥~ In ~ for [2» Q2. Substituting these expressions into (4.5)
and introducing dimensionless variable f2 = l2/Q2 we arrive at the integrals
of the general form
(4.6)
which exhibits n! behavior. Here, the integration goes over 0 < f2 < 1 for
0:> 0 and 1 < [2 < 00 for 0: < o. Finally, one obtains the following large
258 G.P. KORCHEMSKY
order behaviour of perturbative D-function in the large nF limit
CF~
D(Q 2 ) = -;- ~ as n+l[3(1 1-i30,F) nn! (n + 6"11)] . (4.7)
4 '2i3o,F )n n! + '3(
n=O
Notice that the mechanism generating n! terms in D(Q2) is different from

the one produced by instantons. In the latter case n! counts the number of
Feynman diagrams, while the same terms in (4.7) come from a single bubble
diagram. That is the reason why n! divergences of (4.7) are associated with
a contribution of a new object called renorma[on. Two terms in (4.7) scale
at large n as n! and (_l)nn! and therefore the corresponding series have
different convergency properties. These terms originate from two different
regions [2 « Q2 and [2 » Q2 and we shall refer to them as infrared (IR)
and ultraviolet (UV) renormalons, respectively.
Eq. (4.7) is valid only for the large number of light quark flavors, nF » 1
while in QCD this number is not particularly large: nF = 5 for Q2 = M~.
The question arises whether nonleading l/nF corrections could modify the
asymptotic behavior (4.7). To answer this question one invokes a physical
interpretation of (4.5) : a single renormalon chain "renormalizes" the scale
of the coupling constant from external high energy scale Q2 down to gluon
virtuality [2
a s (Q2) L [as (Q2)i30 In(Q2/12)r = a s (l2) . (4.8)

n
Notice that in comparison with (4.5) we replaced fermionic part of the

i3-function by an exact one-loop expression. This is certainly the case in
QED while in QCD the inclusion of nonabelian correction /30 = /3o,G + /3o,F,
the so-called "naive nonabeliazation", implies certain restrictions on the
form of nonleading coefficients Dk,m in (4.4) . Namely, one assumes that
numerical values of these coefficients are dominated by terms,...., i30G , and as
a consequence, (4.7) holds beyond large nF limit after one replac~s i3o,F -+
/30'
To understand the difference between IR and UV renormalons it is
instructive to identify the region of gluon momenta providing a dominant
contribution to (4.5). Examining the momentum integral (4.5) one finds
that the dominant contribution comes in both cases from vicinity of the
saddle points located at
(4.9)
in the case of IR and UV renormalons, respectively. Then, for fixed energy

scale Q2 the position of the saddle points changes as one goes to higher
orders of perturbation theory. One finds that for n -+ 00 the infrared sad-
dle point moves to the origin, lfR -+ 0, producing Borel nonsummable /3~n!
contribution. At the same time, ultraviolet saddle point increases to in-
finity, lDv -+ 00, and its contribution, (-/3o)nn! becomes Borel summable
provided that /30 > O. Thus, IR and UV renormalons probe QCD dynamics
at large and short distances, respectively. The fact that UV renormalons
do not generate any ambiguities in QCD is in agreement with asymptotic
freedom at short distances. It is only in the IR region that we expect nonper-
turbative effects to manifest themselves in perturbative series in the form of
IR renormalon ambiguities. These properties should be compared with the
situation in QED where the large order behavior of the D-function is given
by (4.7) with the QED beta-function /3~ED = /30,F taking negative values.
Repeating analysis one finds that, in contrast with QCD , the ambiguities
of perturbative series in QED are caused by UV renormalons and not by
IR renormalons. Indeed, QED is trivial at large distances and it is only at
short distances that one expects to encounter Landau pole problem.
Using (4.7) and (3.8) it becomes straightforward to perform the Borel
transformation of the D-function. Analysing the properties of B[D](O") we
identify the Borel singularities as
B[DJ( ) - 3CF
0" - 27r 2 - /300"
1 +
CF
37r
[1 5]
(1 + /300")2 + 6(1 + /300") + ...
(4.10)
Here, only contribution of closest to the origin singularities is explicitly

written and remaining singularities are denoted by dots. We conclude that
the leading IR renormalon of the D-function is located at a = 2//30 and
the leading UV renormalon at 0" = - 1//30, Calculating D - function as an
integral over Borel parameter
D(Q2) = fo
00
dO"e- u /a. s (Q2) B[DJ(O") = fo
00
dO"
(A2~~D ) u/3o B[D] (0")
(4.11)
we do not have any difficulties with UV renormalon contribution while the
leading IR renormalon yields an ambiguity in the perturbative definition of
D of the form
(4.12)
This ambiguity indicates that leading nonperturbative effects contribute to

the physical D-function at the level of 1/Q4_power corrections
A4
D(Q2) = DpT(Q2) + c ~~D + 0(1/Q 6 ) . (4.13)
260 G.P. KORCHEMSKY
Here, eAtcn is a nonperturbative scale and 1/Q6 correction corresponds

to the contribution of sub leading IR renormalon located at a = 3/130.
Analysis of IR renormalon allows to determine the form of leading power
corrections but it does not provide any information about physical meaning
of a new scale eAtcn as well as its magnitude. To answer these questions
one needs nonperturbative methods for calculating the D-function. Luckily
enough, the D-function belongs to a special (narrow) class of observables
for which such method exists [4]. It is based on the Operator product ex-
pansion of the product of electromagnetic currents entering the definition of
the polarization operator (4.2). At large negative q2 = _Q2 the interaction
occurs at short distances x 2 1/ Q2. This suggests to expand the product
f'.J
JJ.t(x)Jv(O} at short distances over the set of local composite gauge invari-
ant QCD operators. Substituting this expansion into (4.2) and performing
Fourier integration one finds that in the expression for the polarization op-
erator and as a consequence the D-function the matrix element of local
composite operator is accompanied by the power of high energy scale JQ2
equal to the scaling dimension of the operator. For instance, the contribu-
tion of identity operator scales as 1/ QO and it is associated with perturba-
tive series. The first nontrivial operator with the lowest scaling dimension
is the gluon condensate (OI~FJ.tv(O}FJ.tv(O}IO). Since its scaling dimension
equals 4, the contribution of the gluon condensate to the D-function has
the form of 1/Q4_power correction and it matches exactly the IR renor-
malon contribution (4.13). Comparing two expressions we conclude that
nonperturbative scale parameterizing 1/Q4 corrections in (4.13) is deter-
mined by gluon condensate
(4.14)
Similar identification holds for subleading IR renormalon contribution to

(4.13) while the form of corresponding local operators of sub leading scaling
dimension becomes more complicated. Most importantly, the structure of
power corrections induced by IR renormalon ambiguities turns out to be in
agreement with thorough analysis based on the operator product expansion
[5]. We should notice that the operator product expansion works only for
specific QCD observables like the total cross-section of e+ e- annihilation
and it is not applicable for typical QCD observables like event-shape vari-
ables. In the latter case, IR renormalon becomes a unique theoretical tool
for analysing the power corrections.
Summarizing, IR and UV renormalons provide a new source of n! diver-
gences in QCD. The known structure of Borel singularities in QCD shown
in Fig. 6. This does not mean however that there are no other sources of n!
divergences in QCD. The success of IR renormalon based phenomenology
Instanton-antiinstanton
singularities at cr=4lt. 81t ...
UV renormalons at IR renormalons at
cr=-n/~. n;1,2, ..... cr= n /~. n= 2 , 3 , " .
Figure 6. Known structure of Borel singularities of the Adler D-function
suggests however that IR renormalons provide the leading power correc-

tions.
5. Power corrections to the event shapes in e+ e- -annihilation

Event shape variables represent an important class of QeD observables
and their study provides an information about the fine structure of the
final hadronic states in e+ e- -annihilation. One of the advantage of the
event shapes is that they are infrared finite quantities in QeD and there-
fore they can be calculated in perturbative QeD at large center-of-mass
energy 8. However, comparing two-loop perturbative QeD predictions with
the experimental data one discovers [6] that nonperturbative hadronization
corrections are anomalously large even at high energy 8 = Ml Moreover,
in many cases hadronization corrections scale with the energy as 1/ v's
which should be compared with 1/8 2 behavior for the D-function. A novel
feature of the event shapes is that the operator product expansion is not
applicable in this case and the structure of the leading power corrections
becomes different [7].
Let us consider the simplest event shape - the thrust variable T as a
case example. For a given final state in e+ e- annihilation shown in Fig. 7
the thrust is defined as follows
(5.1 )
Here, the sum goes over final state hadrons carrying space-like momenta Pi
in the center-of-mass frame, Li Pi = 0, and the unit vector ii (ii 2 = 1) is
chosen in such a way that the r.h.s. of (5,1) takes a maximally possible value.
262 G.P. KORCHEMSKY
p.
1
Figure 7. Definition of the thrust axis
Preferable direction defined by the vector n is called the thrust axis. Thus
defined thrust variable takes the values T ~ 1 and has a simple physical
meaning - it measures how narrow are hadronic jets in the final state. For
instance, for infinitely narrow pencil-like jets the thrust axis goes along
jet momenta leading to T = 1, while for spherically symmetric isotropic
configuration of jets the thrust approaches the value T = 0.64.
Analysing the thrust in perturbative QCD one applies quark-hadron
duality and calculates the thrust on par tonic level by replacing hadron
momenta in the definition (5.1) by momenta of outgoing quarks and glu-
ons. Such approximation is valid up to hadronization corrections which are
expected to decrease with the center-of-mass energy.
e-
9 +
e+ q
Figure 8. Gluon radiation correction to e+ e- -+ qq.
At the Born level of perturbative QCD calculation, the final state in

e+ e- annihilation consists of a pair of quark and antiquark moving back-
to-back with momenta fiq and Pit = -fiq leading to
(5.2)
for n '" fiq. Perturbative Qs(s) correction to (5.2) comes from emission of
additional gluon into the final state as shown in Fig. 8. Calculating its
contribution to the thrust is becomes convenient to apply Sudakov decom-
position of gluon momentum
k JL = QP~ + f3p~ + ki
with 0 < a, {3 < 1 and ki = a{3s . Then, one finds for thrust
1- T = min{a,{3} (5.3)
Calculating the mean value (1 - T) one has to average this expression with
probability to emit gluon into the final state
(1 - T) ~ as(s) CF
7r
r dad{3 rs dki 5(sa{3 - ki)
l
io a{3 io
1 {'IS dk.l
4a s (s) CF- (5.4)
7r v's io
with CF = (N; - 1)/(2Nc ) = 4/3 being quark Casimir operator. Going
through two-loop order calculation one finds "state-of-the-art" perturbative
prediction for the mean value of thrust
However, comparing this prediction with existing data (see Fig. 9) one
observes a significant deviation which slowly decreases with the energy as
(5.5)
To explain the origin of this power correction we shall go through the

analysis of large order n! perturbative corrections. Namely, one starts with
the large nF limit and then applies "naive nonabeliazation" to identify the
leading IR renormalon contribution to the mean value of thrust. In this way
one finds that the Feynman diagrams generating n! divergences have the
form of Fig. 8 with a bare gluon propagator replaced by a single renormalon
chain. Similar to (4.5) their contribution is given by
(1 - T)renormalon =
(5.6)
and it amounts to setting the scale of the coupling constant to the transverse
momentum of emitted gluon. Performing integration in (5.6) one finds the
n! behavior of perturbative series as
( T renormalon -_ 4 CF
1 -) as-(s) 1 ~ n () '( {3 )n
- ~ ~ as s n. 2 0 . (5.7)
7r Y s n
264 G.P. KORCHEMSKY
Going through Borel transformation one finds

1
B[1 - T](a) ,..., 1 (5.8)
f30a - 2
and identifies the leading IR renormalon singularity as a = 1/{2f3o). Com-

paring (5.8) with analogous expression for the D-function (4.10) we find
that the leading IR renormalon for the mean value of thrust located closer
to the origin than the one for the D-function. As a consequence, nonper-
turbative corrections induced by IR renormalon ambiguity become stronger
for the thrust than for the total cross-section of e+ e- annihilation
(1 - T) nonPT"'" e- 1/(2/3oCl: s (s» ,..., CT AQCD (5.9)

.;s
with cTAQCD being a nonperturbative scale. We observe that IR renormalon
induced leading power correction to the mean value of thrust agrees well
with expected form of hadronization corrections (5.5) as far as the energy
dependence is concerned. IR renormalons do not provide however the value
of the scale cTAQCD. Performing a global fit of the data for (1 - T) with
1/.;s power correction included one finds
cTAQCD ~ 1 GeV. (5.10)
As can be seen from Fig. 9, I/Q power correction significantly improves

QCD prediction for (1 - T) and the quality of the QCD fit turns out to be
better than the one produced by applying phenomenological Monte-Carlo
event generators.
6. Concluding remarks
Infrared Renormalons offer a simple and efficient method for estimating
power corrections to perturbative QCD predictions. The number of their
successful phenomenological applications is constantly growing and IR renor-
malon analysis is now transforming into one of the standard QCD tools.
One should bear in mind however that IR renormalons allow to parame-
terize leading power corrections but they do not tell us much about their
physical origin. To actually calculate the power corrections and understand
the physics behind them we are still missing nonperturbative methods. We
would like to conclude these lecture with the list of open questions:
- What is a physical meaning of the scale parameterizing 1/Q corrections
to the thrust?
- Does the same scale appear in different event shapes and therefore is
universal?
0.20
o TASSO
<> Mark II
o AMY
0.15 " ALEPH
x OPAL
O(as 2 )
O(as 2)+ l/E cm
.......... JETSET
E-< 0.10 HERWIG
I
.....
--------
0.05
0.00
o 20 40 60 80 100
Ecm (GeV)
Figure 9. Mean value of thrust: comparison of the data with the QeD predictions
- Can one construct the event shapes that are not contaminated by large
l/Q hadronization corrections?
- Are there are Borel singularities different from renormalons and in-
stantons?
- Does large nF limit correctly describe the leading IR renormalon sin-
gularities?
These problems are waiting for their solution and are the subject of ac-
tive research in QCD. Due to space-time limitations we were not able to
go through numerous interesting applications of IR renormalons in these
lectures. Interested readers are advised to consult the review papers [3, 6,
7,8,9, 10, 11J for more details.
I would like to thank the Organizers for an excellent organization of the

School and for kind hospitality at Cargese.
References
1. 't Hooft, G. (1977), in Whys of subnuclear physics, Proc. Int. School, Erice, Italy,
ed. A. Zichichi, Plenum, New York.
2. Zinn-Justin, J. (1981), Phys. Rept. 70, 109-167.
3. Beneke, M. (1999), Phys. Rept. 317, 1-142 [hep-ph/9807443].
266 G.P. KORCHEMSKY
4. Shifman, M.A., Vainshtein A.1. and Zakharov V.1. (1979) Nucl. Phys. B 147, 385.
5. Mueller, A.H. (1992), in QCD: 20 Years Later, Aachen, Germany, ed. P.M. Zerwas
and H.A. Kastrup, World Scientific, Singapore.
6. Webber, B.R. (1994), in Hadronic aspects of collider physics, Zuoz Summer School,
p.49 [hep-ph/9411384].
7. Korchemsky, G.P. and Sterman G. (1995), in QCD and high energy hadronic inter-
actions, 30th Rencontres de Moriond, ed. J. Tran Thanh Van, Editions Frontieres
[hep-ph/9505391].
8. Sachrajda, C.T. (1996), Nucl. Phys. Proc. Suppl. 47, 100 [hep-lat/9509085].
9. Akhoury, R. and Zakharov V.I. (1997) Nucl. Phys. Proc. Suppl. 54A, 217 [hep-
ph/9610492].
10. Braun, V.M. (1997), in Beyond the standard model 5, Balholm, Norway, pp.341-346
[hep-ph/9708386].
11. Dokshitzer, Yu.L. (1999), in 11th Rencontres de Blois: Frontiers of Matter , Chateau
de Blois, France, [hep-ph/9911299].
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD
MODEL
FABIO ZWIRNER
Istituto Nazionale di Fisica Nucleare, Sezione di Padova
and
Dipartimento di Fisica, Universitd di Padova,
Via Marzolo 8, 1-35131 Padova, Italy
e-mail: zwirner~pd.infn.it
These three lectures are a short, non-technical review of the physics of

realistic supersymmetric models, for a mixed audience of young theorists
and experimentalists, and should be used as an invitation to the study of
the rich available literature, not as a substitute for it. The first lecture is
a general introduction to the subject: after a critical look at the Standard
Model of strong and electroweak interactions, it reviews the motivations
for low-energy supersymmetry and the recipe for the construction of su-
persymmetric Lagrangians. The second lecture describes the Minimal Su-
persymmetric extension of the Standard Model and reviews the 'standard'
expectations for the phenomenology of supersymmetric Higgs bosons and
of supersymmetric particles at colliders. The third lecture outlines some
possible steps towards a more fundamental theory, covering topics such as
supersymmetric grand-unification and, at a simplified level, spontaneous
supersymmetry-breaking, superstring unification, large extra dimensions.
LECTURE I: GENERAL INTRODUCTION
1. A critical look at the Standard Model
All confirmed experimental data in particle physics are in agreement with

the Standard Model (SM) of strong and electroweak interactions. The only
ingredient of the SM that has escaped detection so far is the elusive Higgs
boson. Its search is ongoing at LEP (the present lower bound on the SM
Higgs mass is around 106 GeV [1]) and, whether or not evidence is found
at LEP, will continue at the Tevatron and at the LHC. In the absence
267
J-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 267-327 .
268 FABIO ZWIRNER
of direct and unambiguous experimental evidence, the discussion of any

possible physics that goes beyond the 8M is subject to strong theoretical
prejudice. To enable the reader to share the origin of this prejudice, we
begin these lectures with a critical look at the 8M, trying to identify its
virtues and its unanswered questions. Extensions of the 8M such as low-
energy supersymmetry are required to answer some of the latter, but they
should not spoil the former: as we shall see, this is not an easy task!
1.1. THE BUILDING BLOCKS OF THE SM
The theoretical pillar of the 8M is local gauge invariance with respect to

the gauge group
GSM == SU(3)c X SU(2)L x U(l)y. (1)
Gauge invariance completely determines the spin-1 particle content of the

8M: the gluons G~ (A = 1, ... ,8), associated with the strong interactions
and characterized by the coupling constant gs; the bosons (I = 1,2,3) W£
and B JJ , mediating the electroweak interactions with coupling constants g
and g', respectively, and corresponding, when rearranged into appropriate
linear combinations, to the photon, and to the W± and ZO bosons. Gauge
invariance also fixes completely the Yang-Mills part of the Lagrangian,
including the cubic and quartic self-interactions among the non-abelian
gauge bosons, depicted in Fig. 1:
[, y M = _!GJJvAG A
4 JJV
-!4 WJJvIWIJJV _ !BJJV B
4 JJV , (2)
where
(3)
B JJv oJJB v - ovBJJ ,
and!ABC and !IJK are the fully antisymmetric SU(3) and SU(2) structure
constants, respectively.
The spin- ~ particle content of the 8M consists in three generations
of quarks and leptons, whose transformation properties under GSM are
summarized below
qaL == ( ~:~ ) f"V (3,2, +1/6), laL == ( VaL )

eaL
f"V (1,2, -1/2) ,
(4)
UaR f"V (3,1, +2/3), daR f"V (3,1, -1/3), eaR f"V (1,1, -1).
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 269
v
v v
v v
v
Figure 1. Pictorial representation of the cubic and quartic self-interactions among the
non-abelian gauge bosons of the 8M, denoted by the generic symbol V.
In eq. (4), a = 1,2,3 is a generation index, and the weak hypercharge Y

is normalized according to Q = T3L + Y, where Q is the electric charge
and T3L the third component of the weak isospin. We have used left- and
right-handed chiral projections, defined by PL,R = (1 ±,5)/2, and SU(3)c
and SU(2)L indices have been left implicit. Notice the absence of right-
handed neutrinos liaR. Given the quantum number assignments of eq. (4),
gauge invariance completely determines the interactions between fermions
and gauge bosons, depicted in Fig. 2:
(5)
where
D J.l = !l
uJ.l -
•
~gs
cJ.l"
A \A
-
.
~g
I
WIJ.l 2
T
- .
~g
'B J.l Y • (6)
All the fermions are denoted by the collective symbol W == (qL, uR, dR,
1L, eR ) a= 1 ,2,3, and the symbols ,\ A, r;
and Y, appearing in the covariant
derivative Dp., stand for the hermitean generators of the different G SM
factors in the representation defined by eq. (4).
Figure 2. Pictorial representation of the interactions between the 8M fermions, denoted

by the generic symbol J, and gauge bosons, denoted by the generic symbol V.
270 FABIO ZWIRNER
The last but not the least important ingredient of the 8M is a complex
spin-O SU(2)-doublet, the so-called Higgs field:
</J == ( ~: ) '" (1,2, +1/2), (7)
which appears not only in the part of the Lagrangian containing the spin-O
fields gauge interactions and self-interactions, depicted in Fig. 3:
LS = (D~</J)t (D~</J) - V, (8)
V = j.L 2 </J t </J+).. (</Jt</J) 2 , (9)

but also in the one containing the Yukawa couplings, depicted in Fig. 4:
where for notational convenience we have introduced ¢ == (iq 2 </J*) = (cpo* -

cp-)T", (1,2, -1/2), and hU , hD and hE are arbitrary 3 x 3 complex ma-
trices in generation space.
v , s
s ,
,, , ,,
gl,g ",,' ,,
~::""
:'gI2,g2
V ,,
,,
"" . s ,,
,,
v s
s s
,, ,
,, ,,
,, , ,,
,,
, , ,,
, ,x, , A.
, ,, ,,
,,
, , ,,
,, ,
S S
Figure 3. Pictorial representation of the gauge interactions and self-interactions of the
8M spin-O fields, denoted by the generic symbol S.
s ---------
f
Figure 4. Pictorial representation of the Yukawa interactions between the 8M fermions,
denoted by the generic symbol j, and the 8M spin-O fields, denoted by the generic symbol
s.
1.2. SPONTANEOUS BREAKING OF THE GAUGE SYMMETRY IN

THESM
The part of the SM Lagrangian involving the spin-O field ¢ is instrumental

to describe two crucial physical phenomena, to be discussed in turn in the
present and in the following subsection. £s in eq. (8) is a tool to describe
the spontaneous breaking of the local gauge symmetry SU(2)L x U(l)y
down to U(1)Q, with the associated mass generation for the physical W±
and zO bosons.
Choosing A. > 0 and J.L2 < 0, the classical potential of the SM, eq. (9),
is minimized for
(11)
Correspondingly, non-vanishing masses are generated for the W± and ZO

bosons
(12)
whereas the photon remains massless, as dictated by the residual gauge

invariance with respect to U(1)Q. Pictorially, one can understand the origin
of the gauge boson masses by looking at the second graph of Fig. 3 and by
replacing the scalar fields with their constant vacuum expectation values
(VEVs).
For a modern discussion of possible extensions of the SM, an important
information is the fact that the SM description of the spontaneous breaking
of the electroweak gauge symmetry has been tested to an impressive level
of precision at LEP, at the Tevatron and in other experiments at lower
energies. Many observable quantities that are sensitive to the SM radiative
corrections and, potentially, also to new physics, have been measured with
272 FABIO ZWIRNER
high accuracy. The picture that emerges is summarized in Fig. 5 [2], which
includes recent direct measurements of mw from LEP2 [3] and Tevatron [4]
with better than per-mille accuracy, and is in excellent agreement with the
SM predictions!. Now that the top quark mass has been directly measured
with very good precision by the Tevatron experiments [4], the data are
sufficiently precise to be sensitive to the mild, logarithmic dependence of
the SM radiative corrections on the Higgs boson mass mH, and favour
values of mH close to the present experimental upper bound, as can be
seen in Fig. 6 [2]. Combining direct and indirect bounds, one obtains a
99% c.l. upper bound on the SM Higgs mass of roughly 300 GeV. The
most important message of electroweak precision tests, however, concerns
possible physics beyond the SM: only very delicate deviations from the
SM predictions are allowed. This is a very strong constraint on theorists'
imagination, and allows to discard several extensions or modifications of
the SM proposed in the past. For recent reviews of electroweak precision
tests and of their interpretation, within and beyond the SM, see e.g. [6].
1.3. EXPLICIT BREAKING OF THE FLAVOUR SYMMETRY IN THE SM
The second important physical phenomenon where the SM spin-O field plays
a crucial role is the explicit breaking of the global flavour symmetry, via the
Yukawa couplings of eq. (10). In the absence of LYuk, the SM Lagrangian
has a huge [U(3)j5 global symmetry, corresponding to unitary transforma-
tions in generation space for the five irreducible fermionic representations
of the gauge group, eq. (4). Indeed, since the U(l)y subgroup is gauged,
the additional global symmetry is more precisely [SU(3)]5 x [U(1)]4. The
flavour symmetry implies that gauge interactions do not distinguish among
the three generations of quarks and leptons. In the real word, this symmetry
must be broken, since we observe a complicated pattern of masses, mixing
angles and phases for the SM fermions. The role of LYuk in the SM is pre-
cisely the explicit breaking of the flavour symmetry, down to the [U(1)]4
associated with the total baryon number B and the individual lepton num-
bers (Le, Lp., Lr). These correspond to accidental global symmetries of the
SM, in agreement with the experimental bounds on baryon- and lepton-
number non-conserving processes. The fact that B and L == Le + Lf.' + Lr
are violated at the quantum level by non-perturbative effects and its con-
sequences are discussed in other lectures at this School [7].
The Yukawa part of the SM Lagrangian realizes the explicit breaking of
the flavour symmetry in a very special way. On the one hand, the breaking
is very strong, as one can see by staring at the observed pattern [8] of the 9
lFor completeness, we should add that a recent determination ofthe weak charge from
parity violation in Cs atoms [5], not included in Fig. 5, has a pull of about +2.3.
Stanford 1999
Measurement Pull
rTl z IGeVI 91 .1871 ± 0.0021 .08

lzlGeVI 2.4944 i: 0.0024 ' .56
(I
<J harJr Inb] 41.544 :':.0.037 1.75
R(l 20.768;t 0.024 1.16
Ao. e 0.01701 ± 0.00095 .80
!b
. (U4i,n , n.Oot')1 ;:; ,
.,~
;':'"
~:..."
A.'! O.1 42f1 .. !. a.(j('j<l'i ·· 13)1

sin20~~1 0.2321 ± 0.0010 .60
,») C~~':\ 'j nn.S5U G OJ:-·:{.~ .. G?
.-
Rb 0.21642 ± 0.00073 .81

Rr: 0.lG74 ..:: 0.0038 -1.27
A(;,h
!b
0.0988 ± 0.0020 -2.20
AO <' 0.0692 i: 0.0037 .. 1.23
' 1:>
1\ 0.9 '11 .;. 0.025 -.95
A<.. O.6:iO ± 0.026 -1 .46
SI
'n"d"pt
dt 0.23099 ± 0.00026 .. 1.95
"
~~in(·(l~.o:! () 2~?fiSt n.OO~~1
.,: ...:
1 . : :j
""
mw [GeVl 80.448 ± 0.062 1.02

m, [GeV) 174.3 ± 5.1 .22
L'lahad(m Z
(5) )
0.02804 ± 0.00065 -.05
-3 -2 -1 0 1 2 3
Figure 5. Results of a recent fit to precision electroweak observables, compared with

the corresponding 8M predictions.
fermion masses and of the 4 parameters (3 mixing angles and 1 CP-violating

phase) appearing in the Cabibbo-Kobayashi-Maskawa (CKM) matrix. On
the other hand, the only source of flavour violation in the SM Lagrangian
is precisely the CKM matrix, controlling the weak charged-current interac-
tions. In the SM, all the tree-level flavour-changing-neutral-current (FCNC)
274 FABIO ZWIRNER
Figure 6. Results of a fit of the electroweak precision data to the Higgs boson mass mH
within the SM.
couplings, i.e. those of the photon, of the ZO and of the physical Higgs bo-
son, are flavour-diagonal. FCNC processes are induced only by loop effects,
controlled by the CKM matrix and sufficiently suppressed to guarantee
agreement with experimental data on flavour physics.
To conclude this subsection, we recall some important constraints from
flavour physics [8] that are passed with flying colours by the SM, but are
very severe censors of its possible modifications. As for 'quark flavour',
t1mK = (3.491 ± 0.009) x 10- 12 MeV, (13)
IEKI = (2.28 ± 0.01) x 10- 3 , (14)
t1mBd = (3.10 ± 0.11) x 10- 13 GeV, (15)
BR(B --t X s ,) = (3.14 ± 0.48) x 10- 4 , (16)
dn <9X 10-26 e em , (17)
... ,
with a meaning of the symbols that should be obvious to the particle physi-
cists in the audience. As for 'lepton flavour',
BR(J.L --t e,) < 1.2 x 10- 11 , (18)
BR(KL -+ J.Le) < 3.3 x 10- 11 , (19)

de <4 X 10- 27 ecm, (20)
Finally, as for total baryon and lepton number,
t({3{3)ov (76 Ge)

1/ 2
> 1. 1 X 1025 yrs , (21)
r(p -+ e+7r°) > 2.9 x 1033 yrs, (22)

r(p -+ K+lI) > 6.7 x 1032 yrs, (23)
As we shall see, all these constraints can be violated if we extend the SM!
1.4. THE SM AS AN EFFECTIVE THEORY AND ITS PROBLEMS
Despite the remarkable achievements described in the previous subsections,

no physicist believes that the SM is really the ultimate theory of elementary
particles, since, among the other things, it has about 20 arbitrary param-
eters, which may seem too many for a fundamental theory, and it leaves
several unanswered questions, for example some concerning unification and
flavour. The unification problem is related to the gauge interactions, whose
pattern of groups and representations is complicated and arbitrary. Why
should there be three different factors in the gauge group, with the as-
sociated coupling constants taking the values they do? Why should the
fermions transform according to such an odd choice of chiral representa-
tions of SU(2)L x U(l)y, so that parity is violated in weak interactions? The
flavour problem has to do with the Yukawa interactions of the SM, which
introduce several arbitrary parameters into the model. There is no expla-
nation for the existence of three fermion generations with the same gauge
quantum numbers, nor for the complicated observed pattern of masses,
mixing angles and phases.
The previous arguments suggest to go beyond the SM, but there is
more: it is quite obvious that the SM must be extended! Among the 'hard'
arguments supporting the previous statement, the strongest one is the fact
that the SM does not include a quantum theory of gravitational interac-
tions. Immediately after comes the fact that some of the SM couplings are
not asymptotically free, making it almost surely inconsistent as a formal
Quantum Field Theory.
This does not give us direct information on the form of the required
SM extensions, but brings along an important conceptual implication: the
SM should be seen as an effective field theory [9], valid up to some physical
276 FABIO ZWIRNER
cut-off scale A. Assuming that the 8M correctly identifies the degrees of

freedom at the electroweak scale (this may not be true, for example, in the
case of the 8M Higgs field), the basic rule of the game is to write down the
most general local Lagrangian compatible with the 8M symmetries [i.e. the
SU(3) x SU(2) x U(l) gauge symmetry and the Poincare symmetry], scal-
ing all dimensionful couplings by appropriate powers of A. The resulting
dimensionless coefficients are then to be interpreted as parameters, which
can be either fitted to experimental data or (if one is able to do so) theo-
retically determined from the fundamental theory replacing the 8M at the
scale A. Very schematically (and omitting all coefficients and most indices,
as well as many theoretical subtleties):
Leff A4 + A2q,2
+ (Dq,)2 + 'l1 -fb'l1 + p2 + 'l1'l1q, + q,4
'l1'l1 q, 2 + 'l1UJ'V 'l1 F J'V
__
+ A A
'l1'l1'l1'l1 q,2p2
+ A2 +1\2
+ ... , (24)
where 'l1 stands for the generic quark or lepton field, q, for the 8M Higgs
field, P for the field strength of the 8M gauge fields, and D for the gauge-
covariant derivative. The first line of eq. (24) contains two operators carry-
ing positive powers of A, a cosmological constant term, proportional to A4 ,
and a scalar mass term, proportional to A2. Barring for the moment the
discussion of the cosmological constant term, which becomes relevant only
when the model is coupled to gravity, it is important to observe that no
quantum 8M symmetry is recovered by setting to zero the coefficient of the
scalar mass term. On the contrary, the 8M gauge invariance forbids fermion
mass terms ofthe form A'l1'l1 and gauge-boson mass terms. The second line
of eq. (24) contains operators with no power-like dependence on A, but
only a milder, logarithmic dependence, due to infrared renormalization ef-
fects between the cut-off scale A and the electroweak scale. The operators
of dimension d :s; 4 exhibit two remarkable properties: all those allowed
by the symmetries are actually present in the 8M; both baryon number
and the individual lepton numbers are automatically conserved. The third
and fourth line of eq. (24) are the starting point of an expansion in inverse
powers of A, containing infinitely many terms. For energies and field VEVs
much smaller than A, the effects of these operators are suppressed, and the
physically most interesting ones are those that violate some accidental sym-
metries of the d :s; 4 operators. For example, a d = 5 operator of the form
'l1'l1q,2 can generate a L-violating Majorana neutrino mass of order C pl / A,
where G p1 / 2 ~ 300 GeV is the Fermi scale (see [10, 11]). Some of the d = 6
four-fermion operators can be associated with flavour-changing neutral cur-
rents (FCNC) or with baryon- and lepton-number-violating processes such
as proton decay, and so on.
At this point, a question naturally emerges: where is the cut-off scale
A, at which the expansion of eq. (24) loses validity and the SM must be
replaced by a more fundamental theory? Two extreme but plausible answers
can be given:
(I) A is not much below the Planck scale, Mp == (87rG N )-1/2 ~ 2.4 X
1018 GeV, as roughly suggested by a nalve extrapolation of the mea-
sured strength of the fundamental interactions, including the gravita-
tional ones.
(II) A is not much above the Fermi scale, as suggested by the idea that
new physics must be associated with electroweak symmetry breaking.
In the absence of an explicit realization at a fundamental level, each of
the above answers can be heavily criticized. The criticism of (I) has to
do with the existence of the 'quadratically divergent' scalar mass opera-
tor, which becomes more and more 'unnatural' as A increases above the
electroweak scale [12]. On general theoretical grounds, we would expect for
such operator a coefficient of order 1, but experimentally we need a strongly
suppressed coefficient, of order G F1 / A2. However, after taking into account
quantum corrections, this coefficient can be conceptually decomposed into
the sum of two separate contributions, controlled by the physics below and
above the cut-off scale, respectively. Answer (I) would then require a sub-
tle (malicious?) conspiracy between low-energy and high-energy physics,
ensuring the desired fine-tuning. The criticism of (II) has to do instead
with the d > 4 operators: in order to sufficiently suppress the coefficients
of the dangerous operators associated with proton decay, FCNC, etc., the
new physics at the cut-off scale A must have quite non-trivial properties!
On purely dimensional grounds, the effective operators associated with elec-
troweak precision tests would suggest A ,(: 103 Ge V, those associated with
FCNC would suggest A ,(: 106 GeV, those associated with proton decay
would suggest A ,(: 1015 GeV, and so on for many other examples (how-
ever, we should keep in mind that the coefficients of these operators may
be suppressed by loop factors or by symmetries of the underlying funda-
mental theory). As we shall see, this is a potential problem also for the
supersymmetric extensions of the SM discussed in the present lectures.
At the moment, answer (I) is not very popular in the physics commu-
nity, since we do not have the slightest idea on how the required conspiracy
could possibly work at the fundamental level. Conceptually, such a possibil-
ity can be theoretically tested in an ultraviolet-finite Theory of Everything:
278 FABIO ZWIRNER
as daring as it may sound, with the advent and the continuing development
of string theories and their generalizations, we may not be very far from
the implementation of the first quantitative tests. More concretely, such a
possibility can be experimentally tested in the near future, via the search
for the Higgs boson at LEP, at the Tevatron and at the LHC. A clear pic-
ture of the implications of (I) is given in Fig. 7, which shows, for various
possible choices of A in the 8M, the values Higgs boson mass allowed by
the following two requirements [13]:
- The 8M effective potential should not develop, besides the minimum
corresponding to the experimental value of the electroweak scale, other
minima with lower energy and much larger value of the Higgs field. In
first approximation, this amounts to requiring the 8M effective Higgs
self-coupling, A(Q), not to become negative at any scale Q < A: for a
given value of the top quark mass M t , this sets a lower bound on the
8M Higgs mass mHo
- The 8M effective Higgs self-coupling should not develop a Landau pole
at scales smaller than A: for a given value of M t , this sets an upper
bound on mH. 8uch constraint has a meaning which goes beyond per-
turbation theory, as suggested by the infrared structure of the 8M
renormalization group equation for A(Q) and confirmed by explicit
lattice computations [14].
Fig. 7 includes some recent refinements [15] of the original analysis, such
as two-loop renormalization group equations, optimal scale choice, finite
corrections to the pole top and Higgs masses, etc. For very large cut-off
scales, A = 1016 _10 19 GeV, the results are quite stable and can be sum-
marized as follows: for a top quark mass close to 175 GeV, as measured
at the Tevatron collider, the only allowed range for the 8M Higgs mass is
130 GeV .:s mH .:s 200 GeV. This means that, even in the absence of a direct
discovery of new physics beyond the 8M, answer (I) could be falsified by
LEP, the Tevatron and the LHC in two possible ways: either by discovering
a 8M-like Higgs boson lighter than 130 GeV, or by excluding a 8M-like
Higgs boson in the 130-200 GeV range!
Answer (II), instead, gives rise to a well-known conceptual bifurcation:
(lIa) In the description of electroweak symmetry breaking, the elementary
8M Higgs scalar is replaced by some fermion condensate, induced by a
new strong interaction near the Fermi scale. This includes old and more
recent variants of the so-called technicolor models [16] ('extended',
'walking', 'non-commuting', ... ). The stringent phenomenological con-
straints on technicolor models coming from electroweak precision data
and from flavour physics will be mentioned later. On the theoretical
side, technicolor remains quite an appealing idea, still waiting for a
satisfactory and calculable model. The lack of substantial theoretical
600
:I: 400
:::a
200
Figure 7. Bounds in the (Mt, mH) plane, for various choices of A.
progress in this field, however, may be due to the technical difficulties

of dealing with intrinsically non-perturbative phenomena.
(lIb) The SM is embedded in a model with softly broken global supersym-
metry, and supersymmetry-breaking mass splittings between the SM
particles and their superpartners are of the order of the electroweak
scale. This approach, generically denoted as low-energy supersymme-
try, ensures the absence of field-dependent quadratic divergences, and
makes it 'technically' natural that there exists scalar masses much
smaller than the cut-off scale. Moreover, a minimal and calculable
model is naturally singled out, the MSSM. This is the approach that
will be followed in the rest of these lectures.
1.5. A MORE CONCRETE LOOK AT THE NATURALNESS PROBLEM
To understand better the motivations for low-energy supersymmetry, al-

ready outlined in the previous paragraph, we take now a more concrete
look at the naturalness problem. Such problem arises whenever we insist,
as in the SM, on the presence of an elementary Higgs field in the Lagrangian
to describe the breaking of the electroweak symmetry, and we want to ex-
trapolate the model to a scale A much larger than the Fermi scale. The
280 FABIO ZWIRNER
tree-level potential of the SM is characterized by a mass parameter J.L2 and

by a dimensionless quartic coupling A. One combination of these two pa-
rameters, essentially J.L2 / A, is fixed by fitting the VEV v of the SM Higgs
field to the measured value of the Fermi constant, defining the scale of elec-
troweak symmetry breaking. The squared mass m1I of the physical Higgs
particle, proportional to J.L2, or, equivalently, to AV 2, is instead a free pa-
rameter of the SM. While the lower bound on the Higgs mass comes from
experiment, arguments based on perturbative unitarity and triviality sug-
gest that self-consistency of the SM is broken unless
mH < 0(1 TeV). (25)

This is hard to reconcile, from the effective field theory point of view,
with the fact that, already at one-loop, there are quadratically divergent
contributions to the Higgs boson mass, as can be checked by performing an
explicit calculation with a naIve cut-off regularization in momentum space.
The question then arises: how can the Higgs boson mass be of the order of
the electroweak scale and not of the order of the physical ultraviolet cutoff
of the theory?
The problem outlined above is generic for theories containing elemen-
tary spin-O fields. For example, consider a model with a complex spin-O field
of mass mB and a two-component fermion of mass mF, with a Yukawa cou-
pling AF and a quartic scalar coupling AB. The one-loop corrections to the
boson mass include two quadratically divergent contributions of opposite
sign, one involving a fermion loop and controlled by the Yukawa coupling
AF, the other one involving a scalar loop and controlled by the four-point
coupling AB, and have the form
(26)
where A is the ultraviolet cutoff, the minus sign comes from the fermion
loop, and the dots stand for less divergent terms. The situation is radically
different in the case of the loop corrections to the fermion mass, the latter
being protected by a chiral symmetry in the limit mF -+ O. The one-
loop diagram correcting the fermion mass is logarithmically divergent and
proportional to the fermion mass, giving
(27)
Therefore, the fermion mass can be naturally small. In the case of the scalar
mass, what we need to make it naturally small is a symmetry relating bosons
and fermions, and enforcing the vanishing of the coefficient of A2 in (26),
not only at one loop but also at higher orders: the only known candidate
is supersymmetry.
1.6. OTHER MOTIVATIONS FOR SUPERSYMMETRY

Before starting the discussion of low-energy supersymmetry, it is appro-
priate to recall that there are other theoretical motivations to consider
supersymmetry:
- it is the most general symmetry of the S-matrix consistent with a non-
trivial relativistic quantum field theory [17];
- it is an interesting laboratory for the analytical study of the non-
perturbative regime of non-trivial four-dimensional quantum field the-
ories [18];
- it seems to play an important role for the consistency of superstrings
[18], candidate unified theories of all interactions, including the gravi-
tational ones.
However, only the naturalness problem requires the existence of supersym-
metric particles with masses within the Te V scale, making low-energy su-
persymmetry testable at present and forthcoming colliders, and a suitable
subject for a School where experimental physicists are such a large part of
the audience.
1.7. THE COSMOLOGICAL CONSTANT PROBLEM
To conclude this section, we should mention another naturalness problems

of the SM, in some sense analogous to the one discussed for the mass term
of the elementary spin-O field: not only it is very important, but also it
will playa role in the discussion of supersymmetry-breaking in the third
lecture. It is the cosmological constant problem (for reviews and references,
see e.g. [19]), which originates from the fact that a constant contribution to
the vacuum energy, such as the A4 term in eq. (24), is also allowed by the
SM symmetries, and becomes physically relevant when the SM is coupled
to gravity. From the present observations on the expansion rate of the
universe, we can derive the following bound on the cosmological constant
term (for a review of recent data, see e.g. [20]):
C- 1
A cosm '"< Mp
F
rv
10- 4 eV. (28)
From the above equation, it is immediate to realize that the cosmological

constant problem is by far the toughest naturalness problem in fundamental
physics. To make the cosmological constant natural, one would need to
modify gravitational interactions at distance scales of order 1 mm. Such a
possibility is not excluded (see, e.g., [21] for some theoretical speculations
along these lines), but no convincing and self-consistent model has emerged
so far.
282 FABIO ZWIRNER
2. Generalities on supersymmetry (simplified)

The formulation and the perturbative properties of supersymmetric field
theories are described in many excellent textbooks and reviews (see, e.g.,
refs. [22]-[26]). This section summarizes, in a non-technical way, the main
ingredients that playa role in the construction of supersymmetric exten-
sions of the SM at the electroweak scale. The non-expert reader is urged to
consult the pedagogical literature on this subject for a systematic and self-
contained presentation. Concerning the notation, we shall use rather freely
either two-component spinors in the conventions of [22] or four-component
spinors in the conventions of [27].
2.1. SUPERSYMMETRY ALGEBRA AND SUPERFIELDS
In 'ordinary' relativistic quantum field theories, such as the SM, we are

used to deal with the following two types of symmetries:
1. Space-time symmetries, such as the four-dimensional Poincare sym-
metry. The six generators Mp.v of the Lorentz transformations and the
four generators Pp. of the translations obey the known commutation
relations.
2. Internal symmetries (global or gauged), described by some generators
T a , obeying the commutation relations of the corresponding Lie alge-
bra.
The generators of space-time symmetries and the generators of internal
symmetries commute: [Ta,Pp.] = [Ta,Mp.v] = o. As a consequence, internal
symmetries relate particles with the same mass and the same spin. To
introduce symmetries that relate particles of different spin, without spoiling
other important properties of relativistic quantum field theories, one must
generalize this framework and move to supersymmetric field theories.
Supersymmetric field theories [28] are based on the supersymmetry al-
gebra [29], a graded extension of the Poincare algebra, obtained from the
latter by adding some generators of fermionic character, obeying anticom-
mutation relations. We limit ourselves here to the case of simple (N = 1)
supersymmetry in d = 4 space-time dimensions. Most realistic models are
based on this case, which allows for matter fields transforming in chiral rep-
resentations of the gauge group. Realistic models with extended (N > 1)
supersymmetry are more difficult to construct and will not be discussed
here, even if their special field-theoretical properties may justify dedicated
investigations [18]. The basic anticommutation relations of the N = 1 su-
persymmetry algebra are, in two-component notation:
(29)
The supersymmetry generators Q and Q have spin-1/2, as could be seen by

looking at their commutation relations with the generators MI'II of angular
momentum. Also, they commute both with the generators PI' of space-
time translations and with the generators Ta of possible (global and/or
local) internal symmetries. This implies that particles sitting in the same
irreducible representations of supersymmetry have spins differing by 1/2,
but the same internal quantum numbers and, as long as supersymmetry is
unbroken, the same mass.
The most convenient way to classify the representations of supersym-
metry and to construct actions invariant under supersymmetry is to make
use of superfields. The superspace is defined via the generalized coordinates
z = (x, 0, 0), where x are the usual space-time coordinates, and 0 and 0 are
two-component anticommuting coordinates. A superfield is a function in
superspace, and can be expanded in ordinary fields as follows
4>(z) f(x) + OX(x) + OX(x) + (J(Jm(x) + (J(Jn(x)

(30)
+ (JC1I'OV Il (X) + (J(J(J>'(x) + OO(J1jJ(x) + (JOO(Jd(x).
To obtain irreducible linear representations of supersymmetry, suitable
constraints must be imposed on the generic superfield. The two types of
supermultiplets used in the construction of globally supersymmetric exten-
sions of the 8M are the chiral and the vector superfields. In a convenient
basis for the superspace coordinates, chiral superfields have the following
simple power expansion:
¢(x,O) = cp(x) + Y201jJ(x) + (J(JF(x) , (31)
where cp is a complex spin-O field, 1jJ a left-handed two-component spinor

and F a complex scalar, corresponding to an auxiliary non-propagating
field. In the Wess-Zumino gauge, vector superfields can be expanded as
- - - - 1-
V(x, 0, 0) = -OC1 Il OVIl (X) + iO(J(J>'(x) - iOOO>'(x) + 20000D(x), (32)
where VI' is a real spin-1 field, >. and Xare two-component spinors of op-
posite chiralities, and D is a real scalar auxiliary field. From the vector
superfield we can construct the supersymmetric generalization of the gauge
field strength, a chiral superfield given by
284 FABIO ZWIRNER
2.2. RENORMALIZABLE LAGRANGIANS WITH N = 1 GLOBAL

SUPERSYMMETRY
With the previous superfields, we can easily construct the most general
supersymmetric, gauge invariant, renormalizable Lagrangian, exploiting the
fact that under a supersymmetry transformation the highest component of
a superfield transforms into a total derivative. In the case of a simple gauge
group G, to which we associate the vector superfields V == vaTa (where
Ta are the hermitean generators of G) and the gauge coupling constant g,
the result is:
In the above equation, w(<p) is a gauge-invariant polynomial of degree

three in the chiral superfields <pi, called superpotential. Working with four-
component Majorana spinors, and eliminating the auxiliary fields via their
algebraic equations of motion, we obtain: 0
+ [iV2g1/JiAa(Ta<p)i + hoC.] - ~ [a!:;cpi1[?1f? + h.C.]- V(<p, <pt), (35)
where the scalar potential reads
(36)
Notice how supersymmetry brings along a unification of couplings. In

ordinary theories, such as the SM, one may introduce three different types
of dimensionless couplings: gauge couplings, Yukawa couplings and quartic
scalar couplings. Supersymmetric theories allow only for two different types
of couplings, gauge couplings and superpotential couplings, and the dimen-
sionless couplings appearing in the scalar potential are related to these.
2.3. HOW SUPERSYMMETRY MAY SOLVE THE NATURALNESS

PROBLEM
One of the main features of supersymmetric theories is their milder ultravio-

let behaviour, summarized by the so-called 'non-renormalization theorems'
[30]. For example, there is no independent renormalization of the super-
potential parameters at any finite order in perturbation theory. A related
property is the absence of field-dependent quadratic divergences, as long as
there are no anomalous U(I) factors in the gauge group. We shall now use
this property to give an intuitive explanation of how supersymmetry may
help [31] in the solution of the naturalness problem of the SM.
Another way of looking at the naturalness problem of the SM is to
consider its one-loop effective potential, which contains a quadratically di-
vergent contribution proportional to
(37)
where the sum is over the various field-dependent mass eigenvalues mr(cp),
with weights accounting for the number of degrees of freedom and the
statistics of particles of different spin k In the SM, Str M2 depends on
the Higgs field, and induces a quadratically divergent contribution to the
Higgs squared mass, already identified as the source of the naturalness prob-
lem. A possible solution of the problem may be provided by N = 1 global
supersymmetry. For unbroken N = 1 global supersymmetry, Str M2 is
identically vanishing, due to the fermion-boson degeneracy within super-
symmetric multiplets. The vanishing of Str M2 persists if global super-
symmetry is spontaneously broken and there are no anomalous U(I) fac-
tors [32]. Indeed, to solve the naturalness problem of the SM one could
allow for harmless, field-independent quadratically divergent contribution
to the effective potential: this is actually used to classify the so-called soft
supersymmetry-breaking terms [33], to be discussed later. With typical
mass splittings b..m within the MSSM supermultiplets, the field-dependent
logarithmic divergences in the effective action induce corrections to the
Higgs mass parameter which are at most O(b..m 2 ): the hierarchy is then
stable if b..m ;S 1 TeV.
LECTURE II: PHENOMENOLOGICAL SUPERSYMMETRY
3. The Minimal Supersymmetric Standard Model (MSSM)

We shall now describe the two basic building blocks of the MSSM La-
grangian (for reviews, see e.g. refs. [27, 34]).
3.1. SUPERSYMMETRIC PART OF THE LAGRANGIAN
We are now ready to identify the minimal renormalizable Lagrangian with

global N = 1 supersymmetry that extends the SM one [35].
If we keep G == SU(3)c X SU(2)L x U(I)y as the gauge group, the
spin-l fields of the SM are just replaced by vector superfields. The theory
286 FABIO ZWIRNER
contains then some new spin- ~ Majorana particles, called 'gauginos': the
SU(3) 'gluinos' g, the SU(2) 'winos' W, and the U(1) 'bino' E.
Similarly, the spin- ~ matter fields of the SM are replaced by the corre-
sponding chiral superfields, including, as new degrees of freedom, a com-
plex spin-O field for each quark or lepton chirality state: the 'squarks'
ih == (ilL dL)T, UR, dR and the 'sleptons' [L == (VL h)T, eR, in three
generations as their fermionic superpartners. Remembering that chiral su-
perfields contain left-handed spinors, for each generation we shall introduce
the superfields Q, L, UC, DC and EC, whose fermionic components are qL,
lL, (UC)L == (URY, (dC)L == (dR)C and (eC)L == (eR)C, respectively, where the
superscript c denotes charge conjugation.
Finally, we must introduce additional multiplets containing the spin-O
degrees of freedom necessary for the Higgs mechanism. To give masses to
all quarks and leptons, to cancel gauge anomalies and to avoid a massless
fermion of charge ±1, we must introduce at least two Higgs doublet chiral
supermultiplets
HI == ( :! )
rv (1,2, -1/2), H2 == ( ~i )
rv (1,2, +1/2). (38)
They contain, in addition to the spin-O fields (Hf , HI) and (Ht, Hg), de-
noted here with the same symbols of the corresponding superfields with-
out any risk of confusion, also the associated spinor fields (iIf, iII) and
(fIt Jig), the so-called 'higgsinos'.
With the chiral superfields introduced above, the most general gauge
invariant and renormalizable superpotential is
(39)
In the previous formula, generation indices are understood, but we should

keep in mind that the couplings J.L', (h u , hD , hE) and (A, A', A") are tensors
with one, two and three generation indices, respectively. The first line of
eq. (39) contains only terms which conserve the total baryon and lepton
numbers, B and L, whereas the terms in the second line obey the selection
rule !:iB = 0, I!:iLI = 1, and the ones in the third line !:iL = 0, I!:iB I = 1.
The simultaneous presence of the terms in the second and in the third line
would be phenomenologically unacceptable: for example, there could be
superfast proton decay mediated by the exchange of a squark.
The usual way out from this phenomenological embarrassment is the

assumption of a discrete, multiplicative symmetry called R-parity, defined
as
R = (_1)2s+3B+L , (40)
where S is the spin quantum number. In practice, the R-parity assignments
are R = + 1 for all ordinary particles (quarks, leptons, gauge and Higgs
bosons), R = -1 for their supersymmetric partners (squarks, sleptons,
gauginos and higgsinos).
3.2. SOFT SUPERSYMMETRY-BREAKING TERMS
The choice of the gauge group and of the chiral superfield content, and
the requirement of an exact R-parity, are enough to specify the form of
the globally supersymmetric Lagrangian £susy which extends the SM one.
Notice that, even if the number of particles has more than doubled, we have
so far one parameter less than in the SM, since an independent quartic
Higgs coupling is not permitted by supersymmetry. However, one could
show that the perturbative vacua of the theory described by £SUSy do
not break supersymmetry. Therefore, this cannot be the whole story: we
know that supersymmetry is broken in Nature, since we do not observe, for
example, scalar partners of the electron degenerate in mass with it.
The problem of spontaneous supersymmetry breaking will be briefly dis-
cussed in the third lecture. To parametrize the phenomenology at the elec-
troweak scale, the MSSM Lagrangian is obtained [36] by adding to £susy a
collection £SOFT of explicit but soft supersymmetry-breaking terms, which
preserve the good ultraviolet properties of supersymmetric theories. In gen-
eral, £SOFT contains [33] mass terms for scalar fields and gauginos, as well
as a restricted set of scalar interaction terms proportional to the corre-
sponding superpotential couplings. Representing the squark and slept on
fields as vectors in generation space, the generic structure of the MSSM
soft terms is
-£SOFT = qtm~q + uctm~uc + dctmbdc + ftmil + ectm~ec
(41)
It is useful at this point to count the parameters appearing in eq. (41):
mr (i = Q, U, D, L, E) are 3 x 3 hermitian matrices, Aa (a = U, D, E) are
288 FABIO ZWIRNER
3 x 3 complex matrices, MA (A = 1,2,3) and m~ are complex numbers,

m~l and m~2 are real numbers, for a total of 109 real parameters. Four
of them, however, can be eliminated by non-trivial field redefinitions (one
of which can be used instead to remove the phase from the superpotential
mass parameter /-L), so we end up with 105 new parameters with respect to
the SM.
The disappointingly large number of free parameters is not the only
problem connected with a generic choice of CSOFT. For generic values of
these parameters there can be serious phenomenological problems with
charge and/or color breaking vacua and, as we shall discuss in a mo-
ment, with flavour-changing neutral currents and with new sources of CP-
violation. At the level of the MSSM, the most common tentative solution
to these problems is to assume that the MSSM can be extrapolated to
an energy scale M much larger than the weak scale, where the following
universal boundary conditions can be assigned:
2 1
-2
mQ = mu
- 2
= .. . =mo· , (42)
Au = AD = AE = Ao . 1, (43)
(44)
corresponding to universal masses and cubic couplings for squarks and slep-
tons and to universal gaugino masses. Given these assignments, the corre-
sponding parameters of the effective Lagrangian at the weak scale can be
determined by computing, within the MSSM, the appropriate quantum cor-
rections, whose leading effects can be encoded in a set of renormalization
group equations. With this procedure, one can show that the constraints
from flavour physics and from charge/color breaking are easily fulfilled.
However, the previous assumptions (or alternative ones, provided that they
also lead to realistic phenomenology) must be justified within a specific
model of supersymmetry-breaking: we shall come back to this point in the
third lecture.
The introduction of soft terms, combined with some suitable assumption
such as universality, allows us to obtain a realistic extension of the SM at
the weak scale. We shall describe soon the mass spectrum of the MSSM and
the collider bounds on the plethora of new particles it contains. Before doing
that, however, we would like to rate the performance of the MSSM with
respect to some indirect constraints, those coming from flavour physics and
from electroweak precision measurements, that in general put very severe
constraints on possible new physics at nearby scales.
3.3. MSSM VS. FLAVOUR PHYSICS
Since the early days of supersymmetric phenomenology, it was realized

[36, 37, 38] that, allowing for non-universal soft supersymmetry-breaking
terms, the latter would be subject to very stringent constraints from FCNC
and CP violation. An example is the decay f..L -+ e" subject to the strong ex-
perimental bound [8] BR(f..L -+ e,) < 1.2 x 10- 11 . Off-diagonal slepton mass
terms in generation space, denoted here with the generic symbol 6m 2 , would
contribute to the above decay at the one-loop level, via diagrams involving
virtual sleptons and gauginos, and the previous limit roughly translates into
6m2 /mf < 10- 3 -10- 6 , if one assumes gaugino masses of the order of the
average slept on mass mi (a quite complicated parametrization is needed to
by looking at the KO-}(O, BO-Eo systems, at b -+ s,

formulate the bound more precisely). Similar constraints can be obtained
transitions, at the
electric dipole moment of the neutron, and at other flavour-changing or
CP-violating phenomena. It is important to recall that all these bounds are
naturally respected by the strict MSSM, where the only non-universality
in the squark and slepton mass terms is the one induced by the renor-
malization group evolution from the cut-off scale M to the electroweak
scale. However, the same bounds represent quite non-trivial requirements
on extensions of the MSSM, such as supersymmetric grand-unified theo-
ries (SUSY GUTs) and string effective supergravities, since in general one
expects non-universal contributions to the soft supersymmetry-breaking
masses. Various mechanisms that could enforce the desired amount of uni-
versality, or, alternatively, a sufficient suppression of FCNC and CP viola-
tion without universality, have been discussed in the literature. For reviews
of the theoretical and phenomenological aspects of supersymmetric flavour
physics, see e.g. [39].
Moving to more general considerations, the flavour problem is one of
the key issues in all extensions of the SM, including the supersymmetric
ones. This is due to the fact that in the SM the [SU(3)]5 x [U(1)]4 flavour
symmetry is strongly violated, but all flavour violation is encoded in the
Cabibbo-Kobayashi-Maskawa matrix, so that, thanks to the GIM mecha-
nism, there is natural suppression of all flavour-changing and CP-violating
effects. Any model of new physics must face the flavour challenge, especially
if part of the new physics is close to the electroweak scale. This is certainly
the case of the MSSM, where the supersymmetry-breaking problem and the
flavour problem get mixed. As we shall mention in the third lecture, models
with a light gravitino may naturally explain the absence of non-standard
flavour-violating effects, whereas models with a heavy gravitino may lead
to measurable signals, whose detection would open a window on the physics
at very high scales.
290 FABIO ZWIRNER
Even ensuring that there are no tree-level FCNC, in the MSSM new
contributions to FCNC processes may come from loop diagrams involving
virtual non-standard particles, such as the charged Higgs boson, the stops
and the charginos. Comparison with experiment may then lead to indirect
constraints on the MSSM parameters. Important examples include the fits
to /).mBd and I€KI and to the inclusive b -+ s, rate. IT it were possible to re-
duce the theoretical uncertainties due to perturbative and non-perturbative
effects of the strong interactions, these processes would become a very im-
portant source of indirect limits on the MSSM spectrum.
3.4. MSSM VS. ELECTROWEAK PRECISION TESTS
The impressive amount of data collected in recent years at LEP, at the

Tevatron and elsewhere has confirmed the validity of the SM at an un-
precedented level of precision. Nowadays, when discussing physics beyond
the SM we must take into account that only very delicate deviations from
the SM predictions are still allowed at the presently accessible energies.
In this respect, the MSSM performs very well in comparison with other
candidate models. Thanks to the fact that the soft mass terms are invariant
under the electroweak gauge group, the effects of virtual supersymmetric
particles on observable quantities decouple in the limit of a heavy sparticle
spectrum. Of course, having supersymmetric particle masses much heavier
than the electroweak scale would bring back the hierarchy problem, but
this is a different issue: in practice, decoupling occurs very fast and we do
not need to worry about naturalness in this context. This important MSSM
feature should be contrasted with examples of new physics that do not obey
similar decoupling properties, such as a possible fourth fermion generation,
technicolor, and others.
In the case of a heavy sparticle spectrum, the MSSM predictions for
precision electroweak observables essentially coincide with those of the 8M
for a relatively light Higgs, and the corresponding data do not put very
stringent constraints on the MSSM parameter space. In some special cases,
however, a light sparticle spectrum can give rise to sizeable effects: a large
stop-sbottom splitting, in the presence of relatively small soft masses for
the left-handed components, can give a sizeable positive contribution to
the effective p parameter [40]; loops involving light stops and charginos, or
the top quark and the charged Higgs, may give sizeable corrections to the
effective Zbb vertex, with the possibility of partial cancellations [41]; other
effects related with the threshold behaviour of light charginos in the vector
boson self-energies have been considered [42], but their potential impact
has considerably decreased after the stringent limits on chargino masses
obtained at LEP2 (see later).
SUPERSYMMETRlC EXTENSIONS OF THE STANDARD MODEL 291
In the past, given the large number of MSSM parameters, to perform

global fits it was convenient to organize the data in a model-independent
way, by defining a suitable approximate parametrization, and by compar-
ing the MSSM predictions and the fits to the experimental data in terms
of 3-4 relevant parameters. With the present experimental precision, this
approach looks no longer adequate. In general, the indirect bounds on the
MSSM parameter space from electroweak precision data are weaker than
the bounds obtained from the direct searches. Nevertheless, there are small
regions of the MSSM parameter space where the indirect bounds are the
most stringent ones: to discuss these bounds at the appropriate level of
precision, full MSSM computations are required.
For more details on supersymmetry vs. electroweak precision data, the
reader can consult [43] and the references therein.
3.5. TREE-LEVEL POTENTIAL AND SU(2) x U(l) BREAKING
The tree-level scalar potential associated with the MSSM Lagrangian,

LMSSM = LSUSY + LSOFT, (45)
is a function of all the spin-O fields ofthe model. To discuss SU(2)L x U(1)y
gauge symmetry breaking, it is usually assumed that all squark and slepton
fields have vanishing VEV s, and the attention is restricted to the Higgs
potential:
Vo = mi IHll2 + m~ IH212 + m~ (HIH2 + h.c.)
(46)
where
ml2 =_ J.L 2 + mH
21, - J.L 2 + mH2'
m22 = 2 (47)
and, thanks to the possibility of redefining the phases of the Higgs super-
fields, it is not restrictive to assume that m~ < 0, so that the potential is
minimized for
(48)
For the potential to be bounded from below, we have to require that

S == mi + m~ - 21m~1 ~ O. (49)
In order to get non-vanishing VEVs at the minimum, we must destabilize
the origin in field space:
(50)
292 FABIO ZWIRNER
To minimize the potential, it is convenient to use the auxiliary variables

V2
tan/3 == -, (51)
Vl
so that the minimization conditions assume the simple form

. -2m~ 2 4 m~ - m~ tan /3
sm2/3 = 2 2' v = -=--~---=-=---==------- (52)
ml +m2 92 +9 12 tan2 /3 - 1
With these expressions in our hands, we are now ready to study the MSSM
spectrum.
3.6. THE SPECTRUM: R-EVEN SECTOR
The R-even sector of the MSSM contains, to begin with, all the spin-l and
spin- ~ particles of the SM. The only difference is the fact that the mass
terms for gauge bosons and fermions are now originated by two independent
VEVs. For example, the tree-level expressions for the W and Z masses are
2 2 12
2
mw =
92 (2
Vl +V22) , mz
2
= 9 +2 9 (2
Vl
2)
+v2 . (53)
Quarks of charge Q = 2/3 have tree-level masses proportional to V2, quarks

of charge Q = -1/3 and charged leptons have tree-level masses proportional
to Vl. Neglecting for the moment intergenerational mixing, and considering
for example the third generation,
(54)
where (ht, hb' h r ) are dimensionless Yukawa couplings.
A non-trivial structure arises in the Higgs boson sector, where we have,
to begin with, two complex doublets, Hl and H 2 , amounting to eight real
degrees of freedom. After shifting the fields according to
(55)
and after decoupling the Goldstone bosons, GO = - COS/3Pl +sin/3P2 , G+ =

- cos/3(HI)* + sin/3Ht , G- = (G+)*, we are left with five physical degrees
of freedom. Two of them correspond to a charged (complex) field,
H+ = sin/3(Hl)* + cos /3(Ht), H- = (H+)*, (56)
with tree-level mass
2 2 2
mH± =mW+mA' (57)
where
m~ = -m~ (tan{3 + ta~(3) . (58)
The remaining degrees of freedom correspond to three neutral states. One

of them is CP-odd,
(59)
with mass m~ as in eq. (58). The other two are CP-even, and the corre-
sponding mass eigenstates and eigenvalues are obtained diagonalizing the
mass matrix for Sl and 8 2:
m~ cos 2 {3 + m~ sin2 {3 -(m~+m~)Sin{3COS{3) .

M~= ( (60)
-(m~ + m~) sin{3 cos (3 m~ sin2 {3 + m~ cos 2 {3
The explicit expression for the mass eigenvalues is trivially obtained,
and the corresponding mass eigenstates read, in order of increasing mass,
h = - sina81 + cos aS2, H = cos aS1 + sinaS2, (62)
where the mixing angle a is conventionally chosen such that - ~ ::; a ::; 0
and is given by
sin2a = - sin2{3 (m~ + m~).

mH-m
(63)
h
It is important to notice the tree-level mass relations

2 2 2
mH± =mW+mA' (64)
(65)
which imply
It is also important to realize that, at tree level, all Higgs masses and
couplings can be expressed in terms of two parameters only: for example,
we can choose as independent parameters (mA, tan(3), or (mh' tan(3), or
(mh' mA).
294 FABIO ZWIRNER
3.7. RADIATIVE CORRECTIONS TO THE HIGGS SECTOR
We have seen before that, at the classical level, the MSSM is very predic-
tive in the Higgs sector, thanks to the fact that supersymmetry forbids an
arbitrary quartic term in the scalar potential. In particular, the classical
relation mh < mz is very constraining: if it were rigorously true, it would
allow a decisive test of the MSSM already at LEP2, and today we would be
very close to ruling out the MSSM! However, it is by now well known that
the MSSM Higgs sector, and in particular the upper bound on the lightest
Higgs boson mass, are subject to large, finite radiative corrections, dom-
inated by loops involving the top quark and its supersymmetric partners
[44].
The most important effects show up in the one-loop-corrected mass
matrix for the neutral CP-even states,
(67)
Neglecting D-terms and mixing terms in the stop squark mass matrix, and
assuming a common soft supersymmetry-breaking stop mass mi, the lead-
ing correction is
3 2M4 m2
( ~M2 ) = _ 9 t 10 ~ (68)
R 22 811"2 m~ sin2 (3 g Ml .
From this, it is a simple exercise to derive the one-loop-corrected eigenvalues
mh and mH, as well as the mixing angle 0: associated with the one-loop-
corrected mass matrix (67). The most striking fact in eq. (68) is that the
correction (~Mkh2 is proportional to (Mt /m~). This implies that the
tree-level predictions for mh and mH can be badly violated, and so for the
related inequalities. The other free parameter in eq. (68) is mb but the
dependence on it is much milder.
Over the years, the original calculations were progressively refined by
the inclusion of: mixing effects in the stop sector, resummation of the lead-
ing logarithms via the renormalization group, momentum dependence of
the self-energies, loops of other MSSM particles, the most important two-
loop corrections. The state of the art of the theoretical calculations has been
recently summarized in [45]. For the present value of the top quark mass,
M t ~ 175 Ge V, an average stop mass of 1 Te V and arbitrary stop mixing,
the upper bound on mh is approximately 125 GeV. It is perhaps worth men-
tioning an implicit assumption lying behind the derivation of such upper
bound: non-renormalizable operators, suppressed by inverse power of some
cut-off scale A, should be negligible; indeed, as we shall mention again in the
third lecture, one can build models with very low scales of supersymmetry
breaking, where this upper bound is strongly violated.
3.B. THE SPECTRUM: R-ODD SECTOR
We now move to the spectrum of the R-odd sector of the MSSM.

The spin-O s-particles are the superpartners of the ordinary quarks and
leptons. Even neglecting inter-generational mixing, there is another kind of
mixing that has to be taken into account. Barring the case of sneutrinos, for
which the corresponding fermion is purely left-handed, the spin-O partners
of left- and right-handed quark and leptons can in general mix, and their
mixing is described by 2 x 2matrices of the form
m2
2 iLL
M-= ( (f = e,u,d), (69)
f 2
m-
hR
where
mtL miL (soft) + miL (D - term) + m},
(70)
mlRR mlR (soft) + mlR (D - term) + m},
m2 _{ mf(Af+ptan(3) f=e,p,T,d,s,b (71)
iLR - mf(A f + pcot(3) f = u,c,t
and the D-term contribution is given by
(72)
In general, therefore, one expects the interaction eigenstates, (iL,IR), to
differ from the mass eigenstates, (it, 12) in order of increasing mass. How-
ever, the amount of L-R mixing is proportional to the mass of the corre-
sponding fermion, and is usually negligible for the first two generations.
Among the spin- ~ sparticles, we find the strongly interacting gluinos,
g, which do not mix with other states and whose mass is an independent
parameter of LSOFT.
The weakly interacting spin- ~ sparticles are two charged and four neu-
tral gaugino-Higgsino mixtures, usually called "charginos" and "neutrali-
nos" , respectively.
The two chargino mass eigenstates, (Xf, X~) in order of increasing mass,
are superpositions of winos W± and Higgsinos iIil' and their mixing is
described by the mass Lagrangian: '
296 FABIO ZWIRNER
where the 2 x 2 mass matrix Me is given by
( M2 V2mwsin,8 ) (74)
V2mw cos,8 J1. '
and is diagonalized by the bi-unitary transformation
m .± 0 )
U*Mevt = ( (75)
oXl m.±
X2
.
Similarly, the mixing between the four neutralino states is described by the
mass Lagrangian
(76)
where (~o) T == (iJ, W3, .ilf, .ilg) and the 4 x 4 neutralino mass matrix
reads (c,B == cos,8, s,B == sin,8, Cw == cos Ow, Sw == sin Ow)
o -mzc,Bsw mzs,Bsw
M2 mzc,Bcw -mzs,Bcw
(77)
mzc,Bcw o -J1.
-mzs,Bcw -J1. o
and is diagonalized by the unitary transformation
N* MNNt = diag (m.o

Xl
m·o
X2
m·o
Xs
m.o)
X4
. (78)
Summarizing, the masses and couplings of the two charginos and of the
four neutralinos are characterized by four parameters: the gaugino masses
M2 and Ml (which can be related by a universality assumption, as already
mentioned and further discussed in the third lecture), the superpotential
Higgs mass J1. and tan,8. It should be noted that the lightest neutralino mass
eigenstate, X~, is the favourite candidate for being the Lightest Supersym-
metric Particle (LSP) in the MSSM spectrum. An alternative candidate is
the sneutrino ZIT) but it is actually the LSP of the MSSM for a much smaller
range of parameter space. In general, the lightest neutralino turns out to
be a mixture of the four interaction eigenstates
(79)
The case of a pure photino, X~ = ;y, which was assumed for simplicity
in some old phenomenological analyses, would correspond to the special
combination (Nll,N12,N13,N14) = (sinBw,cosBw,O,O), but there is no

theoretical reason to prefer it.
4. 'Standard' SUSY-Higgs and SUSY phenomenology at collid-

ers
Collider searches for MSSM Higgs bosons and supersymmetric particles
have been extensively discussed in other lectures at this School [3, 4, 46] so
we can afford to be very brief here.
4.1. HIGGS BOSONS SEARCHES
Supersymmetric Higgs bosons have been intensively searched for at LEP,

which in 1999 has delivered about 250 pb- 1 at ...;s = 192..;- 202 GeV. LEP
searches are based on two complementary processes: e+ e- -+ hZ, whose
cross-section is proportional to sin2 (,B - a), and e+e- -+ hA, whose cross-
section is proportional to cos 2 (,B-a). Taking into account that no significant
excesses with respect to the expected background have been reported, the
combination of these two processes should allow to establish, both for h
and for A, an absolute lower bound of the order of 85 Ge V or so, for typical
values of the parameters controlling the radiative corrections [1]. With the
energy and the luminosity of the past years, the Tevatron collider was not
very sensitive to the MSSM Higgs bosons: the present limits on the charged
Higgs mass from top decays [8] are significant only for values of tan,B at
the boundaries of the preferred range 1 < tan,B < Mt!M b . Unfortunately,
even by further raising the energy by a few more GeV in the year 2000,
LEP will not be able to explore completely the parameter space of the
MSSM Higgs sector. In any event, the search for SUSY Higgs bosons will
continue at the upgraded Tevatron and at the LHC. A recent study [47] has
shown that, by looking at a variety of different channels and with sufficient
luminosity, the upgraded Tevatron can significantly expand the region of
parameter space explored by LEP. As for the LHC, the first studies (see,
e.g., [48] and references therein), which focused on the simplified case of
heavy supersymmetric particles, have been considerably improved by the
computation of the most important MSSM corrections to the relevant pro-
duction processes, by the inclusion of possible Higgs decays into pairs of
lighter supersymmetric particles, and by more accurate experimental simu-
lations (see e.g. [49J and references therein). A complete no-lose theorem is
not available, but it seems quite plausible that, if the MSSM is correct, at
least part of its Higgs sector will not escape detection at the LHC. A more
complete exploration of the MSSM Higgs sector could then be pursued at
some high-energy linear e+ e- collider, of the type currently under study
[46].
298 FABIO ZWIRNER
4.2. SPARTICLE SEARCHES
The general framework of supersymmetry is so flexible that it is very diffi-

cult to give a unified description of the searches for supersymmetric parti-
cles. In the following, we shall briefly review the present bounds (no signal of
supersymmetry has been observed yet!) and the future discovery potential,
organizing the discussion around the most important machines contributing
to these searches.
Unless otherwise stated, we shall work under the assumption that there
are no other light states besides those of the MSSM. As we shall discuss
in the third lecture, this is not always the case: the presence of a light
gravitino in the spectrum may lead to a different phenomenology (here and
there we shall anticipate some comments on the light gravitino case).
We shall also assume exact R-parity conservation. Then:
- supersymmetric (R-odd) particles are produced in pairs: single produc-
tion in reactions initiated by ordinary (R-even) particles would violate
R-parity;
- supersymmetric (R-odd) particles always decay into final states involv-
ing an odd number of supersymmetric (R-odd) particles;
- the lightest supersymmetric particle (LSP) is absolutely stable.
If the LSP is neutral and weakly interacting (typical candidates encoun-
tered in model-building are the lightest neutralino or one of the sneutri-
nos in heavy gravitino models, and the gravitino itself in light gravitino
models), then it is a possible candidate for dark matter [11]. In collider
phenomenology, being essentially invisible to the detectors, the LSP can be
characterized by a distinctive missing-energy signature.
Even with the above restrictions, the complex interplay of the depen-
dences of masses, cross-sections and branching ratios on the various param-
eters makes it very difficult to specify simple general limits. Sometimes, one
may choose to combine different searches within the so-called 'constrained
MSSM': this means assuming universal boundary conditions on the soft
masses at the supersymmetric grand-unification scale (to be discussed in
the third lecture), Mu ~ 2 X 1016 GeV, so that the low-energy spectrum
and interactions are essentially described (modulo some subtleties for the
stop sector) by four basic parameters, for example mo, m1/2, !J. and tanj3.
The results of the searches for supersymmetric particles are regularly
updated in the PDG [8].
LEP
LEPl was a solid starting point for deriving very general limits on the
sparticle spectrum. Working on the Z peak, and using both indirect con-
straints from the line shape and dedicated searches, all conceivable decays
of the Z boson into pairs of supersymmetric particles were studied, with

high statistics and controllable backgrounds. As a rule of thumb, this al-
lowed to exclude most supersymmetric particles up to mass values of the
order of mz /2: the only possible exceptions were particles with suppressed
couplings to the Z boson, such as the lightest neutralino X or the lightest
stop iI, for special choices of the corresponding mixing parameters.
At LEP2, the production cross-sections for sparticle pairs are more
model-dependent than at LEP1, but, thanks to the higher energy, much
stronger limits can be obtained [3]. For example, chargino pair production
is controlled by s-channel (r, Z) exchange and by t-channel ve exchange,
with the possibility of (partial) destructive interference in the case of a
light sneutrino. Since chargino decays involve the lightest neutralino, the
mass difference between the lightest chargino and the lightest neutralino is
another important parameter for the searches. Barring special corners of
the parameter space with low acceptance (almost degenerate chargino and
neutralino) or low cross-section (light sneutrino), and given the absence of
a signal over the background, the lower bound on the chargino mass is very
close to the kinematical limit. When the analysis of the 1999 data will be
completed, the typical LEP2 bound on the chargino mass should exceed
100 GeV.
Also the associated production of neutralinos (XX'), of charged sleptons
(1+ 1-) and of stop squarks (il tl) can be used to obtain interesting limits at
LEP2. All these processes occur via s-channel exchange of neutral vector
bosons. In the case of select ron production, there is an important additional
contribution from t-channel neutralino exchange, which may increase the
cross-section substantially. In the constrained MSSM, the combination of
chargino and neutralino searches can be used to set a lower bound of the
order of 30 GeV on the lightest neutralino, and a more stringent bound can
be obtained if one restricts the allowed values for the sneutrino mass and
for tanj3. Typical limits on the charged sleptons are today in the 80-95 GeV
region, depending on the slepton flavour and on some model assumptions,
such as the allowed amount of mass degeneracy between left and right slep-
tons, and between sleptons and the lightest neutralino. One of the reasons
why the sleptons limits are in general weaker than the chargino limits is
the strong p-wave phase space suppression near threshold.
Comparable, sometimes even slightly more stringent limits can be de-
rived for the case of light gravitino, when the lightest MSSM particle is
allowed to decay.
Hadron colliders
Being strongly interacting sparticles, squarks and gluinos are best searched
for at hadron colliders. Both in the heavy and in the light neutralino case,
300 FABIO ZWIRNER
production cross-sections for §§, §q, qq pair-production in pp or pp colli-

sions are relatively model-independent functions of my and mij. As far as
signatures are concerned, one has to distinguish two main possibilities: if
my < mij , then q ~ q§ immediately after production, and the final state
is determined by § decays; if mij < my, then § ~ qq immediately after
production, and the final state is determined by q decays. The first case is
favoured by the constrained MSSM. In old experimental analyses, it was
customary to work under a certain set of assumptions: 1) five or six (qL, qR)
mass-degenerate squark flavours; 2) LSP == i', with mass negligible with re-
spect to mij, my; 3) the dominant decay modes of squarks and gluinos are
the direct ones, § ~ qqi' if my < mij and q ~ qi' if mij < my. The
signals to be looked for are then multijet events with a large amount of
missing transverse momentum. To derive reliable limits, however, one has
to take into account that the above assumptions are in general incorrect.
For example, one can have cascade decays § ~ qqX?#, q'qxt ~ ... and
ij ~ qX?#, q' xt ~ .... The effects of these cascade decays become more
and more important as one moves to higher and higher squark and gluino
masses. Taking all this into account, the present limits from the Tevatron
collider [4] are roughly in the 200 -;- 250 Ge V range. At the LHC (for recent
studies, see e.g. [50]), CMS and ATLAS should be able to explore squark
and gluino masses up to 1-2 TeV, essentially filling the MSSM parameter
space allowed by theoretical prejudices on naturalness.
The searches for charginos and neutralinos at hadron colliders are not
very competitive in the heavy gravitino case. On the other hand, the smaller
backgrounds for the final states with hard photons gives hadron colliders an
advantage in the light gravitino case. In typical models, the present Teva-
tron data give limits on neutralinos and charginos that are more stringent
than the LEP ones.
LECTURE III: TOWARDS A MORE FUNDAMENTAL

THEORY
5. Supersymmetric grand-unification
The basic idea of grand unification is that the gauge interactions as observed
at the presently accessible energies, with the different numerical values of
their coupling constants, are just the remnants of a theory with a single
gauge coupling constant, spontaneously broken at a very high scale. The
simplest possibility is to have a single scale Mu » mz, at which a simple
gauge group G is spontaneously broken down to the SM gauge group:
Mu mz
G --+ GSM == SU(3)c X SU(2)L x U(l)y --+ SU(3)c x U(l)Q
gU (gS, g, g')
(80)
There is a vast literature on grand unification, both with and without su-
persymmetry, and many excellent reviews are available (see e.g. [51]). We
shall limit ourselves here to a qualitative overview of the main differences
between the two cases and to a few comments on some recent developments.
5.1. NON-SUPERSYMMETRlC GRAND-UNIFICATION
The simplest realization of the grand-unification idea is the minimal, non-

supersymmetric SU(5) model of Georgi and Glashow [52] (for a previous
attempt with partial unification, see [53]). The gauge bosons of such model
belong to the adjoint representation of the rank-4 simple group SU(5),
24v: besides the SM gauge bosons, there are 12 additional ones, (X, Y) rv
(3,2, +5/6) and their conjugates (X, Y), of mass Mv. These bosons have
fractional electric charge and carry both baryon and lepton number, flB =
flL = ±1. Each fermion generation is arranged in an anti-fundamental
representation, 5F, and in the antisymmetric product of two fundamentals,
10F. In terms of SM fermions, the two representations decompose as follows:
(81)
Notice that in minimal SU(5), as in the SM, there is no need to introduce a
right-handed neutrino, represented by a left-handed antineutrino V C in the
present conventions. The scalar fields introduced to describe the different
stages of spontaneous symmetry breaking correspond to an adjoint repre-
sentation, 24s, containing 12 Goldstone bosons and 12 additional scalars of
mass Mr;" and an anti-fundamental representation, 5s, containing the SM
Higgs boson and an additional triplet H rv (3,1, +1/3) of mass MH.
The first stage of symmetry breaking is controlled by the VEV of the
24s, of order Mu. The masses Mv, Mr;" MH have model-dependent rela-
tions with M u , but in first approximation we can assume that they are all
of order Mu. The breaking of the SM gauge group at the electroweak scale
is controlled instead by the VEV of the SM Higgs doublet contained in the
5s. The fermions get masses via their Yukawa couplings, of the form
h(lO) . 10F x lOF X 5s, h(5) ·5F x lOF X 5s, (82)
where generation indices have been understood. These Yukawa couplings

cannot give rise to a realistic pattern of fermion masses and mixing (even
302 FABIO ZWIRNER
if some predictions such as the Mb/Mr ratio [54] are intriguingly close to
being correct), but are chosen to keep the model simple.
Non-minimal grand-unified models can be constructed, by enlarging one
or more of the following: the gauge group, the fermion content, the scalar
content. They will not be discussed here. We should mention, however,
the particular appeal of models based on the group 80(10), which can
accommodate, in a single irreducible representation of dimension 16, the
fermions of a full 8M family plus a right-handed neutrino.
One of the most dramatic phenomenological implications of grand uni-
fied models is the possibility of tl.B = tl.L = ±1 nucleon decay, for ex-
ample p -+ e+7r°. There are two types of tree-level Feynman diagrams,
involving three quarks and a lepton on the external lines, that could in-
duce such a process. The first type involves the exchange of virtual (X, Y)
vector bosons on an internal line, and the corresponding rate scales as
r rv 9& / M~j the second type involves the exchange of the scalar Higgs
triplet H, and the corresponding rate scales as r rv h4 /M'k, where h is a
Yukawa coupling. In the case of gauge-mediated nucleon decay, the amount
of model-dependence is small. In first approximation, from the experimen-
tal bound [8] T;:Z:P(P -+ e+7r°) > 2.9 x 1033 yrs, and from the approximate
formula T~h(P -+ e+7r°) rv 1028 ±1 yrs · [Mv(GeV)/2 x 1014 ]4, we can deduce
a stringent lower bound on the grand-unification scale Mu.
The important point is that, from the measured values of two of the
low-energy gauge couplings, we can extract a rather precise prediction for
9u, Mu and the third low-energy gauge coupling. In first approximation,
we can just solve the one-loop renormalization group equations (RGEs) for
the running gauge couplings [55],
d9~ bA 4
dt = 87r29A, (A= 1,2,3j t=logQ), (83)
with the boundary conditions
(84)
where 93 = 9s, 92 =
9 and 91 =
v'5/39'. In the absence of new physics
thresholds between Mu and the scale Q <t: Mu, the RGEs are trivially
solved by
1 1 bA Mu
(85)
9~(Q) = 9~ + 87r21og Q.
The one-loop beta function coefficients appropriate for the 8M are easily
computed [55]:
b2SM = _19 (86)

6'
Starting from three input data at the electroweak scale Q = mz [8], for
example a3, a em and sin2 Ow, we can perform consistency checks of the
grand-unification hypothesis in different models.
In the minimal SU(5) model [52], and indeed in any other model where
eq. (84) holds and the light-particle content is just the SM one (with no
intermediate mass scales between mz and Mu), eqs. (85) and (86) are
incompatible with experimental data. This was first realized by noticing
that the prediction Mu ~ 1014 - 15 GeV, obtained by using as inputs a3
and a em , is incompatible with the limits on nucleon decay. Subsequently,
also the prediction sin2 Ow ~ 0.21 was shown to be in conflict with experi-
mental data [56], and this conflict became more and more significant with
the progressive accumulation of high-quality data from the LEP and Teva-
tron experiments. We shall see in a moment how grand-unification can be
phenomenologically more successful when combined with supersymmetry.
5.2. WHAT CHANGES WITH SUPERSYMMETRY
Some of the problems of non-supersymmetric unification, including those

with proton decay and with the low-energy values of the gauge coupling
constants, may find a natural solution with the incorporation of supersym-
metry. The minimal model of supersymmetric grand unification [57] is based
on SU(5), and is constructed in analogy with the MSSM. Gauge bosons and
matter fermions fall in the same SU(5) representations as in the Georgi-
Glashow model, but are promoted to the corresponding supermultiplets.
The Higgs sector is extended to the following chiral superfields: H(5), H(5)
and :E(24). The VEV of the adjoint scalar, (:E) = V . diag(2, 2, 2, -3, -3)
breaks SU(5) down to the SM gauge group, whereas (H) = (0,0,0,0, V2)
and (H) = (0,0,0,0, Vl) describe the breaking of the electroweak symmetry.
The superpotential is of the form
w = h·lOpxlOpxH+h'·lOpx5pxH+M'HH+).IH:EH+MTr:E2+).2Tr:E3.
(87)
The breaking of SU(5) must preserve supersymmetry and give mass to
the color triplet Higgs bosons, while keeping their doublet partners light.
Looking at the equations of motion for the auxiliary fields, we find that
V'" MI).2 and, in order to keep the Higgs doublets light, M' ~ 3).1 V. The
fine-tuning related to this last condition is at the origin of the so-called
doublet-triplet splitting problem of minimal supersymmetric grand unifi-
cation. The superheavy vector bosons have masses proportional to gu V,
the Higgs triplets in the fundamental and anti-fundamental have masses
proportional to ).1 V, and the Higgs particles in the adjoint have masses
proportional to ).2 V. After decoupling these heavy states, and introducing
304 FABIO ZWIRNER
by hand some soft supersymmetry-breaking mass terms, we are left with

the MSSM as the effective theory at scales Q ~ Mu.
In the leading logarithmic approximation, the predictions of supersym-
metric grand-unification just depend on the MSSM particle content. The
renormalization group equations for the gauge couplings have the form of
eq. (83) and are solved as in eq. (85), the only difference being the one-loop
beta function coefficients. Those appropriate to the MSSM are [58]
(88)
Assuming for simplicity that all supersymmetric particles have masses of

order mz, we obtain [58] Mu ~ 2 X 1016 GeV (which increases the proton
lifetime for gauge-boson-mediated processes beyond the present experimen-
tal limits) and sin2 Ow ~ 0.23. At the time of [58], when data were pointing
towards a significantly smaller value of sin2 Ow, this was considered by some
a potential phenomenological shortcoming of the MSSM. The high degree of
compatibility between data and supersymmetric grand unification became
manifest [56] only later, after improved data on neutrino-nucleon deep in-
elastic scattering were obtained, and was progressively reinforced by the
subsequent LEP and Tevatron data. We should not forget, however, that
unification of the MSSM is not the only solution which can fit the values
of the gauge coupling constants at the electroweak scale extracted from
experiment: for example, non-supersymmetric models with ad hoc light ex-
otic particles or intermediate symmetry-breaking scales could also do the
job. The MSSM, however, stands out as the simplest physically motivated
solution.
In models of supersymmetric grand-unification, including the minimal
one, we still find the conventional mechanisms for proton decay, described
by supersymmetric d = 6 operators. Gauge-boson exchange, however, does
not lead to proton decay at a detectable rate, since the unification mass Mu
is more than one order of magnitude higher than in the non-supersymmetric
case, and the proton lifetime scales as M&. Color-triplet Higgs boson ex-
change could lead to decay modes such as p --t p,+ KO or vf.lK+, but the
corresponding rate would be undetectably small, being proportional to some
Yukawa coupling squared, if the triplet masses are of the order of Mu. How-
ever, as pointed out in [59], supersymmetric models admit a new class of
d = 5 operators which, when dressed by loops of MSSM particles, may lead
to a proton lifetime proportional to I:lm 2 M[; instead of M&, with distinc-
tive decay modes such as p --t K+vw This is indeed the case of minimal
supersymmetric SU(5), whose parameter space is quite severely constrained
by the recent experimental bound r(p --t K+v) > 6.7 X 1032 yrs. However,
the detailed predictions for the decay rates are rather model-dependent,
since they are controlled by superpotential couplings containing two arbi-

trary phases and three independent superheavy masses, and by the details
of the MSSM particle spectrum.
If we want to make the comparison between low-energy data and the
predictions of specific grand-unified models more precise, there are several
factors that should be further taken into account. After the inclusion of
higher-loop corrections and threshold effects, eq. (85) is modified as follows
1 1 bA Mu th l>l
g~(Q) = gb + 87r 210g Q+LlA +LlA (A=I,2,3). (89)
In eq. (89), Ll~ represents the so-called threshold effects, which arise when-
ever the RGE are integrated across a particle threshold [60], and Ll~l
represents the corrections due to two- and higher-loop contributions to
the RGE [61]. Both Ll~ and Ll~l are scheme-dependent, so one should
be careful to compare data and predictions within the same renormaliza-
tion scheme. Ll~ receives contributions both from thresholds around the
electroweak scale (top quark, Higgs boson, and in SUSY-GUTs also the
additional particles of the MSSM spectrum), and from thresholds around
the grand-unification scale (superheavy gauge and Higgs bosons, and in
SUSY-GUTs also their superpartners). Needless to say, these last threshold
effects can be computed only in the framework of a specific grand-unified
model, and typically depend on a number of free parameters. Besides the
effects of gauge couplings, Ll~l must include also the effects of Yukawa
couplings, since, even in the simplest mass-independent renormalization
schemes, gauge and Yukawa couplings mix beyond the one-loop order. In
minimal SU(5) grand unification, and for sensible values of the top and
Higgs masses, all these corrections are small and do not affect substantially
the conclusions derived from the naIve one-loop analysis. This is no longer
the case, however, for supersymmetric grand unification. First of all, one
should notice that the MSSM by itself does not uniquely define a SUSY-
GUT, whereas threshold effects and even the proton lifetime (due to the
new class of diagrams [59J which can be originated in SUSY-GUTs) become
strongly model-dependent. Furthermore, the simplest SUSY-GUT [57], con-
taining only chiral Higgs superfields in the 24, 5 and "5 representations of
SU(5), has a severe problem in accounting for the huge mass splitting be-
tween the SU(2) doublets and the SU(3) triplets sitting together in the 5
and "5 Higgs supermultiplets. Threshold effects are typically larger than in
ordinary GUTs, because of the much larger number of particles in the spec-
trum, and in any given model they depend on several unknown parameters.
Also two-loop effects of Yukawa couplings are quantitatively important in
SUSY-GUTs, since they depend not only on the heavy quark masses, but
also on tanj3: these effects are maximal for tanj3 close to 1 or to Mt!Mb,
306 FABIO ZWIRNER
which correspond to a strongly interacting top or bottom Yukawa cou-

pling. There is no problem of principle in evaluating all these effects, but
they introduce a large amount of model-dependence when we try to push
the comparison between theory and experiment to the level of the present
experimental precision. The conclusion is that, even imagining a further re-
duction in the errors of the experimental determinations of the low-energy
gauge couplings, it is impossible to claim indirect evidence for supersymme-
try and to predict the MSSM spectrum with any significant accuracy. The
only safe statement is that, at the level of precision corresponding to the
naIve one-loop approximation, there is a remarkable consistency between
experimental data and the prediction of supersymmetric grand unification,
with the MSSM R-odd particles roughly at the electroweak scale.
5.3. RENORMALIZATION GROUP EQUATIONS AND IMPLICATIONS
We shall now assume, for the rest of this section, that the MSSM can be
safely extrapolated up to a very large scale M, to be identified, in first ap-
proximation and for the purposes of this lecture, with the grand unification
scale Mu . We shall also assume that, at the scale M, we can assign univer-
sal boundary conditions on the soft terms, in the form of a universal scalar
mass (m~), a universal gaugino mass (ml/2), and a universal cubic scalar
coupling (Ao), all of the order of the electroweak scale. Then the values of
the MSSM parameters at the electroweak scale are strongly correlated by
the corresponding RGEs, whose main features and implications will be now
discussed.
We begin by spelling out in more detail the assumptions on the bound-
ary conditions. We assume for the gaugino masses
(90)
for the soft supersymmetry breaking scalar masses
rh~(M) = rhbc(M) = rhbc(M) = rhHM)
(91)
= rh~c(M) = mk1 (M) = mk2 (M) = ma,
and for the soft supersymmetry-breaking scalar couplings
(92)
We stress that, while (84) and (90) can be justified in models of super-
symmetric grand unification, the universal structure in generation space
of (91) and (92) requires a deeper justification in the underlying theory
of spontaneous supersymmetry breaking. Counting also the supersymmet-
ric Higgs mass J.t(M) = J.to and the supersymmetry-breaking Higgs mixing
term m~(M) = (m~)o, in addition to the gauge and Yukawa couplings we

have in the MSSM five more parameters
(93)
which control the low-energy effective Lagrangian (45).

Some of the RGEs for the MSSM parameters have quite simple approx-
imate solutions. For the gaugino masses, the following equations hold:
dMA bA 2
~ = 81r29AMA, (A = 1,2,3), (94)
and they are immediately solved with the boundary conditions (90), to give
(95)
Numerically, this corresponds to M3 "-' 3 ml/2, M2 "-' 0.85 ml/2, Ml "-'

0.42 ml/2, with possible corrections due to higher-loops and threshold ef-
fects.
Neglecting intergenerational mixing, the one-loop RGE for the top Yukawa
coupling reads [62]
dh t ht ( 8 2 3 2 13 t2 2 1 2)
dt = 81r2 -3 93 - "2 92 - 18 9 + 3ht + "2hb , (96)
with similar equations for the bottom and tau Yukawa couplings, hb and hr-
A close look at the above RGE, combined with the experimental knowledge
of the top and bottom quark masses, can give us important informations.
Consider first the simple case of tan,6 < < Mt!M b • In first approxima-
tion, we can neglect the effects of the (9,9') gauge couplings and of the
(hb' h r ) Yukawa couplings on the running of the top Yukawa coupling, h t .
Then we can immediately realize that the RGE for the top Yukawa coupling,
eq. (96), admits an effective infrared fixed point [63], smaller than in the SM
case [64]. Whatever high value one assigns to the top Yukawa coupling at
the large scale M, the top Yukawa coupling at the electroweak scale never
exceeds a certain maximum value, arax ~ (8/9)as, where at == hU(41r)
and as == 9~ / (41r ). Remembering the tree-level formula for M t , this suggests
the lower bound
tan,6 2: 2. (97)
However, a precise bound can be established only after the inclusion of
the possibly sizeable radiative corrections associated with threshold effects,
both at the unification scale and at the electroweak scale [65], combined
with two-loop RGEs. As a result, values of tanj3 as low as 1.6 may still
308 FABIO ZWIRNER
be acceptable. The bounds of course evaporate if we allow for the possible

existence of new physics thresholds between the electroweak and the grand-
unification scales.
This infrared structure becomes even more interesting if we include the
effects of the bottom-quark Yukawa coupling, so that also large values of
tan f3 can be considered. In this case, the top and bottom Yukawa couplings
admit an effective infrared fixed curve, approximately described by [66]
at + ab ,:s ~as f(at, ab), (98)
where f is a hypergeometric function bounded by 1 <f < 12/7. This

translates into the approximate bound
Ml + ~f3
-·-2-
Ml ~< (200 GeV) 2 . (99)
sm f3 cos
It is remarkable that, for a large range of tan f3 values between 1 and Mt/M b ,
this bound is respected but almost saturated: several theoretical papers
have been written to suggest possible explanations of this empirical obser-
vation, but such a discussion is beyond the aim of the present lectures.
Similar equations can be derived [67] for the soft SUSY-breaking scalar
masses, for the remaining soft SUSY-breaking parameters (At, Ab, AT, m~)
and for the superpotential Higgs mass J.t. Also, the inclusion of the complete
set of Yukawa couplings, including mixing, is straightforward. In general,
the RGE for superpotential couplings and soft SUSY-breaking parameters
have to be solved by numerical methods (or approximate analytical meth-
ods). Exact solutions of the one-loop RGEs can be found for the squark and
slepton masses of the first two generations, for which the Yukawa couplings
are negligible. For example, we get m~, mEre, mbe '" m~ + (5 -;- 8)mi/2'
mi '" m~ + 0.5 mi/2' m~e '" m~ + 0.15 mi/2' with the warning that higher
loops and threshold effects should be included for more accurate predic-
tions.
One of the most attractive features of the MSSM is the possibility of
describing the spontaneous breaking of the electroweak gauge symmetry as
an effect of radiative corrections [68]. Notice that, starting from universal
boundary conditions at the scale Mu, it is possible to explain naturally
why fields carrying colour or electric charge do not acquire non-vanishing
VEVs, whereas the neutral components of the Higgs doublets do. Also,
the electroweak scale gets linked with the scale of the soft supersymmetry-
breaking masses in the MSSM (which remains however an independent
input parameter), and is stable with respect to quantum corrections.
We give here a simplified description of the mechanism, in which the
physical content is transparent, and we comment later on the importance of
a more refined treatment. The starting point are the boundary conditions
on the model parameters at the scale M, summarized by:
gu, (100)
After evolving all the running parameters from the grand-unification scale
M to a low scale Q '" mz, according to the RGEs described in the previous
section, we can consider the RG-improved tree-level potential Vo(Q), which
has the functional form of eq.(46), but is expressed in terms of running
masses and coupling constants evaluated at the scale Q. Vo(Q) will describe
an acceptable breaking of SU(2) x U(I) if the conditions of eqs.(49) and
(50) are satisfied, together with a certain number of conditions for the
absence of charge and colour breaking minima (for recent discussions, see
e.g. [69]), and finally if v2 == v? + vi is of the right magnitude to fit the
observed values of the Wand Z masses, according to eq. (53) . In other
words, the measured values of the weak boson masses set a constraint on
the independent parameters of eq. (100).
A crucial role in the whole process is played by the top quark mass,
since the top quark Yukawa couplings governs the renormalization group
evolution of the mass parameter m~2 :
(101)
where
1:1
L't
2
= mQ + muc
2
+ mH2
2
+ A2t· (102)
For a given set of boundary conditions on the remaining parameters, too

small values of h t are not able to drive B < 0 at scales Q '" mz, so that the
origin remains a minimum and we end up with unbroken SU(2) x U(I); on
the other hand, too large values of h t can either drive S < 0, which would
correspond to a potential Vo(Q) unbounded from below, or violate one of
the conditions for the absence of charge or colour breaking minima.
To conclude the discussion of radiative symmetry breaking, we show
now that in the MSSM (with universal boundary conditions) we expect
Mt
1 < tan/3 < Mb . (103)
The proof relies on the relation, derived from the minimization of Vo( Q):
(104)
310 FABIO ZWIRNER
The boundary conditions at the unification scale are m~(M) = m~(M) =

m5 + J.£5, and the RGE for the difference m~ - m~ reads
d(mi - m~) = _1_ (3h2 Fi

dt 871"2 b b
+ h 2F.
T T
Fi)
_ 3h2
t t , (105)
where Fb and FT are (positive) quantities analogous to Ft. Imagine now

that tanj3 < 1, and remember the tree-level expressions for the top and
bottom masses. The fact that M t ~ Mb then implies h t ~ hb, this in turn
implies that at the scale Q, where the use of Vo(Q) is appropriate, mi > m~.
But eq.(104) then tells us that tanj3 > 1, in contradiction with the starting
assumption. Similarly we can prove that tan j3 < Mt! Mb.
As a final remark, we stress a problem left unsolved by the MSSM de-
scription of radiative symmetry breaking: the scale of the soft terms, which
in turn determines the electroweak scale, is not dynamically determined,
but introduced 'by hand' in the boundary conditions on the mass parame-
ters. To discuss the possible dynamical determination of such a scale, needed
for a fully satisfactory solution of the naturalness problem, we need a theory
of spontaneous supersymmetry breaking.
6. The problem of sllpersymmetry-breaking

To introduce the discussion on spontaneous supersymmetry breaking, it is
useful to assess the achievements of the model with softly broken super-
symmetry discussed so far, i.e. the MSSM embedded in a SUSY-GUT with
universal boundary conditions:
- the hierarchy between the weak scale and the GUT scale is stable, as
long as the SUSY-breaking masses are of the order of the weak scale;
- there is one successful prediction for the gauge couplings at the weak
scale, if there is a 'big desert' between the weak scale and the GUT
scale;
- the breaking of the electroweak gauge symmetry is naturally induced
by quantum corrections, dominated by the large Yukawa coupling of
the top quark;
- electroweak precision tests are easily passed, thanks to the decoupling
properties of supersymmetric particles;
- the precision tests from flavour physics can be also easily passed, after
making the universality assumption on the soft masses.
Accompanying these achievements there are two big question marks. The
first is the observation that no supersymmetric particle has been found yet,
with the experimental bounds constantly improving, cutting more and more
the parameter space allowed by naturalness. The second is the observation
that soft supersymmetry breaking is theoretically unsatisfactory: it cannot
be consistently coupled to gravity, it does not fully solve the hierarchy

problem, and the universality conditions look rather 'ad hoc'. While the
first problem is being addressed by experiments, the second one can be
addressed by moving to the theory of spontaneous supersymmetry breaking.
6.1. SPONTANEOUS SUPERSYMMETRY BREAKING
For the purposes of these lectures, we shall characterize supersymmetry

breaking via an important criterion that follows directly from the basic
anticommutation relation of the supersymmetry algebra, eq. (29), remem-
bering that H = Po is the Hamiltonian. If the Hilbert space has positive
norm, supersymmetry is spontaneously broken if and only if the Hamilto-
nian does not annihilate the vacuum, HIO) =F O. This corresponds in turn to
having a positive vacuum energy, (V) > O. Remembering the structure of
the scalar potential in renormalizable theories with global supersymmetry,
eq. (36), the condition for supersymmetry breaking is then that at least
one of the auxiliary fields of the chiral and vector supermultiplets has a
non-vanishing VEV,
(106)
The unavoidable consequences of the spontaneous breaking of global super-

symmetry are
- The existence of a massless fermion, the goldstino, residing in the super-
fields whose auxiliary fields acquire non-vanishing VEVs (in complete
analogy with the goldstone bosons of ordinary spontaneously broken
continuous global symmetries).
- A positive vacuum energy (we shall describe in a moment what happens
when the coupling to supergravity is introduced).
- Some phenomenologically unacceptable mass relations, such as Str M2 =
o in each separate sector of the spectrum. It should be kept in mind,
however, that such a relation is valid only at the classical level, and
in the absence of non-renormalizable interactions and anomalous U(1)
factors.
The general, 'kinematical' aspects of spontaneous supersymmetry break-
ing are well understood, both in the global [70] and in the local [71] case:
in a N = 1, d = 4 theory with chiral and vector supermultiplets, the or-
der parameters controlling supersymmetry breaking are the VEVs of the
associated auxiliary fields, Fi and n a , which give a positive semi-definite
contribution to the scalar potential. For supersymmetry breaking to be
compatible with a flat space-time background, the inclusion of gravitational
interactions is essential, since in Poincare supergravity the scalar potential
312 FABIO ZWIRNER
reads [72]
(107)
The three terms IIFW, IIDI12 and IIHW are positive-semidefinite, and con-
trolled by the auxiliary fields of the chiral, vector and gravitational super-
multiplets, respectively. The first two terms have different expressions but
identical roles in local and global supersymmetry; the third one, peculiar to
supergravity, has the universal property that (1IHI12) = 3 m~/2M~, where
m3/2 is the mass of the spin-3/2 gravitino (the supersymmetric partner of
the spin-2 graviton).
As will be clear in a moment, to generate phenomenologically acceptable
masses for the supersymmetric partners of ordinary particles, a realistic
model must have
(108)
On the other hand, to satisfy the present bounds on the cosmological con-
stant, a realistic model must also have
(109)
It is then obvious that, when discussing the vacuum energy, the gravita-
tional contribution to the scalar potential must be essentially identical to
the non-gravitational one. However, as we shall see in the following, there
are situations in which gravitational interactions can be neglected when
restricting the attention to the spectrum and the interactions relevant for
present accelerator experiments.
The goldstino G, which provides the ±1/2 helicity components of the
massive gravitino via the super-Higgs mechanism, is determined by
(110)
The mass splittings in the different sectors of the model, denoted here
schematically with a sub-index I, are controlled by
(111)
where >q is the effective coupling of the goldstino supermultiplet to the

sector I. This is true not only at tree level, but also after the inclusion of
quantum corrections, since the latter can be incorporated in a local effective
Lagrangian, which must exhibit the spontaneous nature of supersymmetry
breaking if a full, non-anomalous set of supersymmetric multiplets is kept.
In order for supersymmetry to solve the naturalness problem, it is cus-
tomary to require that the mass splittings among the MSSM states be
(D.m 2 h '" G1/. However, this is not sufficient to fix As or, equivalently,
m3/2 (to an excellent approximation, As = J3 m3/2Mp): according to the
numerical values of the effective couplings AI, different possibilities arise,
to be described in the following paragraphs.
Despite the satisfactory understanding of the 'kinematical' aspects of

spontaneous supersymmetry breaking, what we are still lacking is some
compelling idea about the symmetries and dynamics that control such a
phenomenon in the fundamental theory of Nature, and explain the origin
of the different scales relevant for the problem: D.m 2 , As and Acosm. This
is a very difficult and ambitious problem, and it is not surprising that a
final solution has not been found yet. Several interesting ideas have been
pursued in recent years, but there are still many open problems. We just
mention here some of the existing approaches, referring the reader to the
literature for more details. For a recent review of the possible mechanisms
of supersymmetry breaking, see e.g. [73]. One interesting possibility is that,
in the context of supergravity, the spontaneous breaking of supersymmetry
finds its origin in non-perturbative phenomena, such as gaugino condensa-
tion [74]. Explicit models of this type exist, but they have to rely on some ad
hoc assumptions: being supergravity an effective, non-renormalizable the-
ory, it is difficult to control quantum corrections already at the perturbative
level. Another possibility is spontaneous breaking at the string level, via
coordinate-dependent compactifications [75]. There are however unsolved
problems such as the mechanism for the stabilization of the dilaton VEV
and the generic instability of string vacua with broken supersymmetry and
vanishing cosmological constant with respect to string loop corrections. The
present hope is that some more insight into this mechanism, which may lead
to a non-perturbative formulation of it, could be gained by exploiting the
recently discovered string dualities. A different approach to the study of
spontaneous supersymmetry breaking consists in working at the level of
renormalizable gauge theories with global supersymmetry, and in posing
dynamical questions of more limited scope. Despite the encouraging results
in recent years (for reviews, see e.g. [76]), models of dynamical supersym-
metry breaking at low energy are still quite contrived when one tries to
make them realistic.
Given this state of affairs, in the following we shall give a macroscopic

description of the different scenarios for spontaneous supersymmetry break-
ing, trying to emphasize their generic features and phenomenological im-
plications, and avoiding the discussion of the details of the microscopic
theory.
314 FABIO ZWIRNER
6.2. SUPERGRAVITY MODELS WITH HEAVY GRAVITINO
The first possibility, realized in the so-called hidden-sector supergravity

models, is that the couplings of the goldstino supermultiplet to the MSSM
states are of gravitational strength, AI '" A~/M~. In this case the desired
MSSM spectrum requires As '" G;1/4M~/2 '" 1010 -;- 1011 GeV, and there-
fore m3/2 '" G;1/2. The effective theory at the electroweak scale is obtained
from the underlying supergravity by taking formally the limit Mp ~ 00,
while keeping m3/2 fixed [77]: this gives precisely the MSSM with explic-
itly but softly broken supersymmetry. The states with masses O(m3/2) and
interactions of gravitational strength need not be included in the effective
theory.
In the minimal realization of such a scenario, the superfield content of
the model can be classified in two distinct sectors: the 'observable' sector,
containing the MSSM states, and the 'hidden' sector, containing at least
the gravitational supermultiplet and the goldstino supermultiplet (for defi-
niteness, we assume here that it is a gauge singlet chiral superfield, S). The
two sectors are connected only via non-renormalizable interactions, sup-
pressed by inverse powers of the Planck mass. The scale of supersymmetry
breaking is given by (Fs) '" G;//2 Mp, and the fermionic component of S
is the goldstino G. The gravitino mass is m3/2 '" (Fs)/Mp '" G;l/2, and
the SUSY-breaking mass splittings, both in the observable and in the hid-
den sector, are of the order of the gravitino mass, since they are originated
by tree-level couplings of gravitational strength. In contrast with the case
of renormalizable, global supersymmetry, the supertrace mass sum rule is
in general violated, and the mass scale characterizing such violation is the
gravitino mass.
6.3. SUPERGRAVITY MODELS WITH LIGHT GRAVITINO
The second possibility occurs when the goldstino supermultiplet is coupled

to the MSSM sector by gauge or Yukawa interactions, much stronger than
the gravitational interactions. Taking for example AI '" 1, to get the de-
sired mass splittings one needs As '" G;l/2, giving m3/2 '" Gi/ Mil '"
(few) 10- 5 eV. If there is some weak coupling and the goldstino supermul-
tiplet couples to the MSSM states only via loops, As and m3/2 can increase
by a few orders of magnitude, since the effective couplings AI can be sup-
pressed by numerical factors such as o:/(4tr) and by mass ratios such as
As / M, where M ~ As is some supersymmetry-preserving mass term, pos-
sibly associated with the vacuum expectation value of a standard-model-
singlet scalar field. In this second class of models, gravitational interactions
are relevant only for the discussion of the vacuum energy, and the effec-
tive theory at the electroweak scale can be obtained by taking formally the
naIve limit Mp -r 00, while keeping As constant [78].
A low scale of supersymmetry breaking, As, may be favoured by generic
arguments related with the flavour problem. From the point of view of
the underlying theory with spontaneous supersymmetry breaking, the typ-
ical magnitude of the soft terms in the sfermion sector is A~ / A, where A
is the scale suppressing the corresponding nonrenormalizable operators in
the Kahler potential. If the scale of flavour physics, A ilav , is larger than A,
then we would expect flavour-breaking effects on the soft terms to be sup-
pressed by A/ Ailav, and a phenomenologically acceptable pattern of soft
mass terms could naturally arise. The opposite situation, A ilav ,:s A, would
generically induce unsuppressed flavour violations in the soft terms. These
generic arguments are not conclusive, but may be taken as an additional
motivation to study models where As and A are as low as possible.
A presently popular realization of the light gravitino case is given by the
so-called 'messenger' or 'gauge-mediated' models (for a recent review and
references, see e.g. [79]). In the minimal version of such models, the field
content can be divided into three sectors: an 'observable' sector, containing
the M88M fields; a 'messenger' sector, containing real representations of a
grand-unified gauge group (for example, a 5 +"5 of 8U(5), to be denoted by
M and M, respectively), which interacts with observable sector only via 8M
gauge interactions; a 'secluded' sector, containing at least the gravitational
supermultiplet and the goldstino supermultiplet 8, which has superpoten-
tial interactions with the messenger sector, but is decoupled at tree-level
from the observable sector. If supersymmetry is spontaneously broken on
the vacuum, one expects that the spectrum in the messenger sector is con-
trolled by the combination of supersymmetric mass terms, proportional to
(8), and supersymmetry-breaking masses, proportional to J(Fs). In the
observable sector, supersymmetry breaking masses are generated by loop
diagrams with messenger fields on the internal lines. For example, gaugino
masses are generated at one loop, and have the form
M
A
rv aA J7:FSl. V~(F)
47r (8) \£s/, (112)
whereas universal scalar masses are generated at two loops, and have the
form
2
rno rv 47r
(aA)2
(Fs)
(8)2' (Fs) . (113)
It is easy to identify in the above formulae the effective couplings of the

goldstino supermultiplets to the observable sector, once the effects of loop
diagrams have been included. The nice feature of these models is the fact
that, due to the universal character of gauge interactions, the soft scalar
316 FABIO ZWIRNER
masses in the observable sector are automatically universal. However, be-

cause of a Peccei-Quinn symmetry, neither J.L nor m~ can be generated by
gauge interactions alone, so the minimal messenger model must be compli-
cated with some superpotential interactions in order to become realistic.
Once superpotential interactions are introduced, however, the universality
properties of the scalar mass terms are no longer guaranteed in general.
Moreover, if there is no mixing with the MSSM states, and a conserved
global messenger number can be identified, then we expect a stable mes-
senger, which may give rise to cosmological problems. Both the difficulties
mentioned above can be solved by complicating sufficiently the model, but,
as a result, no unique candidate messenger model is singled out.
In view of the above considerations, a more model-independent ap-
proach to the light gravitino case may be followed (for an extensive dis-
cussion, see e.g. [80]). It consists in writing down an effective theory for
the light multiplets, i.e. the MSSM fields and the gravitino, assuming that
the heavier fields (for example, the messengers, but not necessarily so)
have been integrated out. Such an effective theory has both supersym-
metry and the gauge symmetry linearly realized on the fields, but non-
renormalizable operators are present to encode the low-energy effects of
the underlying dynamics. In this theory, supersymmetry is spontaneously
broken, and masses and couplings can be read off tree-level formulae di-
rectly. The limit of such an approach is the lower amount of predictive
power, but the advantage is the possibility of an efficient parametrization
of the model-independent aspects of the resulting phenomenology. In par-
ticular, the differences with the heavy gravitino case become more and
more important as the supersymmetry-breaking scale As, suppressing the
non-renormalizable operators, gets closer and closer to the weak scale. We
finally remark that an effective theory of this kind is valid only in a limited
energy range, bounded from above by unitarity, which essentially dictates,
besides b..m ;S As, also E ;S A~/ b..m: new (elementary or composite) de-
grees of freedom must be introduced before or near this critical scale to
restore unitarity.
6.4. PHENOMENOLOGY OF SUPERSYMMETRY-BREAKING
Depending on the scale of supersymmetry breaking (or, equivalently, on the

gravitino mass), three broad scenarios for supersymmetric phenomenology
emerge, whose general features will be now described.
Heavy gravitino
This corresponds to m3/2 '" 102 -;- 104 GeV, or As '" 1010 -;- 1011 GeV.
As discussed before, in the heavy gravitino case all polarization states of
the massive gravitino couple with gravitational strength, and the MSSM
with soft terms is an adequate description up to energy scales of order

Mu. The two most distinctive phenomenological features are that non-
renormalizable operators correcting the MSSM are completely negligible
at present accelerator energies, and that the LSP belongs to the MSSM
spectrum.
Light gravitino
This corresponds to m3/2 '" 10- 1 -;- 103 eV, or As '" 104 -;- 106 GeV. In
this case, the ±1/2 helicity components of the gravitino, corresponding to
the would-be goldstino, couple with strength much greater than gravita-
tional, but still smaller than the typical strength of the gauge interactions
or of the Yukawa interactions of heavy fermions. In this case, the new
non-renormalizable interactions, correcting the MSSM and associated with
supersymmetry breaking, are too weak to playa role in the production pro-
cesses of R-odd particles, but may play an important role in their decays.
Also, we can no longer extrapolate the MSSM up to Mu, since tree-level
unitarity is violated at a critical energy Ec '" A~/ D.m, and new (elementary
or composite) degrees of freedom must be introduced before or near this
critical scale to restore unitarity.
An important property controlling the phenomenology of these models,
whose LSP is the gravitino, is the nature of the next-to-lightest super-
symmetric particle (NLSP). If such particle is the lightest neutralino, for
example the photino, the rate of its decay into a photon and a goldstino is
given by [81]
-)
r ( l' -t G, = - 1 --2
MP
. (114)
167l" A~
This is trivially generalized to the case of an arbitrary neutralino, as long as
it has a non-negligible photino component. In this case, the typical signature
of sparticle production and decay is given by photons plus missing energy.
If the NLSP is a sfermion j, for example a stau or a sneutrino, as it may
be the case in some of the messenger models, then it likes to decay into the
corresponding fermion I and a goldstino. In the m f = 0 limit, the decay
rate reads
- -) 1 in}
r ( I-tGI = 167l"A~. (115)
In this case, the phenomenology is characterized by missing energy signals,

as in the standard case of heavy gravitino.
Superlight gravitino
This corresponds to m3/2 '" 10- 6 -;- 10- 2 eV, or As '" 102 -;- 104 GeV.
In this case, the goldstino couplings with the MSSM fields have, at the
318 FABIO ZWIRNER
presently accessible energies, a strength comparable with the gauge cou-

plings. As a result, it is essential to keep track, at energies of the order
of the electroweak scale, of all the leading non-renormalizable interactions
controlled by inverse powers of the supersymmetry-breaking scale. In fact,
as we shall see in a moment, these interactions can now play an important
role in the production processes: we can have not only pair-production of
MSSM sparticles, but also associated production of a gravitino and a MSSM
sparticle, and even pair production of gravitinos. It is also clear that in this
case the effective theory has a very limited range of validity, extending not
much above the electroweak scale.
To conclude the discussion of the superlight gravitino case, we would like
to comment further on an intriguing aspect of its phenomenology. There
may be experiments where the available energy is still insufficient for the
on-shell production of other supersymmetric particles, but nevertheless suf-
ficient to give rise to final states with only gravitinos and ordinary parti-
cles, at measurable rates. As recently discussed in [82], powerful processes
to search for a superlight gravitino G (when the supersymmetric partners
of the Standard Model particles and of the goldstino are above threshold)
are e+e- -+ GG, and qq -+ GG" which would give rise to a distinctive
(photon + missing energy) signal. The first process can be studied at e+ e-
colliders such as LEP or the proposed NLC, the second one at hadron col-
liders such as the Tevatron or the LHC. At hadron colliders, we can also
consider the partonic subprocesses qq ~ GGg, qg -+ qGG, qg -+ qGG
and gg -+ gGG, all contributing to the (jet + missing energy) signal.
In the case of heavy superpartners, all these processes have cross-sections
with a strong, universal power-law dependence on the centre-of-mass en-
ergy and on the scale of supersymmetry breaking, s3 / A~. In the absence of
experimental anomalies, the above processes can be used to establish model-
independent lower bounds on the gravitino mass. From the present LEP
data, we can estimate m3/2 ~ 10- 5 eV, corresponding to As ~ 200 GeV. At
hadron colliders, the analysis is more complicated. In the (,+ h) channel,
there are already some published DO data collected at the Tevatron collider,
from which we can extract As > 210 GeV, or m3/2 > 1.1 X 10-5 eV. We
estimate that, with the presently available luminosity, the Tevatron exper-
iments should be sensitive up to As ~ 265 GeV, or m3/2 ~ 1.7 X 10- 5 eV.
The sensitivity should be slightly higher in the (jet+ lET) channel: our
estimate is As ~ 310 GeV, or m3/2 ~ 2.3 X 10-5 eV. At the LHC, because
of the pp initial state, the most sensitive channel will be (jet+ Ffr), which
should reach As ~ 1.6 TeV, or m3/2 ~ 0.62 X 10-3 eV.
As a final remark, we would like to stress that m3/2 (t-t As) is a funda-
mental free parameter for supersymmetric models, analogous to the Fermi
constant G F for the models of weak interactions, so it is very important to
measure it or at least to bound it from below.
7. String unification and extra dimensions

As discussed in other lectures at this School [18], the most fundamental
unified theories conceived so far are superstring theories. In the pertur-
bative formulation of these theories, point particles, whose propagation in
space-time is associated with a one-dimensional world-line, are replaced by
one-dimensional extended objects, strings, whose propagation in space-time
is described by a two-dimensional world-sheet. This replacement removes
the concept of interaction vertices localized at given space-time points, and
leads to a much softer ultraviolet behaviour of the string scattering ampli-
tudes.
Among the well-established results of superstring theories, all of which
were already known at the end of the eighties, we can mention the following:
- they automatically include quantum (super)gravity, in a way that guar-
antees the absence of ultraviolet divergences, order-by-order in pertur-
bation theory;
- they automatically achieve the unification of all interactions, including
the gravitational one: the only free parameter is the string tension,
which fixes the unit of measure of the massive string excitations, and all
the other scales and parameters are related to VEVs of scalar fields, the
so-called moduli, corresponding to flat directions of the scalar potential;
- in perturbation theory, there are five different consistent theories in d =
10, all with some amount of space-time supersymmetry: two heterotic
theories of closed strings, with gauge group SO(32) or Es x Es and
N = 1 space-time supersymmetry, two type-II theories of closed strings
with N = 2 supersymmetry, one non-chiral (IIA) and the other one
chiral (IIB), and one type-I theory with open and closed strings and
N = 1 supersymmetry;
- after compactifying six out of their ten dimensions, superstring theories
may include, in appropriate limits, some four-dimensional supersym-
metric unified field theories.
Traditionally, the discussion of superstring phenomenology, and in par-
ticular of the unification of all couplings, used to be given in the context of
the perturbative formulation of Es x Es heterotic string models, compact-
ified down to four space-time dimensions. In such a context, the relevant
scales are the following: the string scale, Ms == 0.'-1/2, as defined by the
masses of the string excitations, which can be identified with the scale
where ordinary quantum field theory breaks down and 'stringy' effects take
over; the (phenomenological) four-dimensional Planck scale, Mp, defined
in term of Newton's constant as measured over macroscopic distances; the
320 FABIO ZWIRNER
compactification scale, Me == 1/ R, which for isotropic compactifications

of d extra dimensions can be associated with the average compactification
radius R V~. In the simple case d = 1, Me is linked with the existence
I'V
of Kaluza-Klein excitations of masses Mn = n / R (n = 0, 1, 2, ... ), and

analogous formulae hold for different values of d.
The generic problems of perturbative superstring phenomenology were
soon identified: there were too many perturbatively acceptable supersym-
metric vacua, and there was substantially no control over non-perturbative
phenomena, essential to obtain a realistic model. More specific problems
were arising from the key relations among the relevant scales, the unified
string coupling constant gs and the unified field-theory coupling constant
gu:
Ms = gu . Mp , gs = gu . R3 . M~ . (116)
Since gu, as extrapolated from low-energy data, turns out to be of or-
der unity, to avoid a strong coupling regime one was forced to assume
Ms rv Me rv Mp. Then one had to face a small unification problem: the
value of Ms would turn out to be roughly a factor of 20 higher than the
extrapolated value of the supersymmetric grand-unification scale Mu. This
was pointing towards large threshold effects, which in turn would natu-
rally ask for a large compactification radius, which in turn would lead to a
strongly coupled string [83]. Also in the framework of the supersymmetry-
breaking problem, perturbative string constructions were suggesting that,
to have supersymmetry-breaking mass splittings of the order of the weak
scale, one would also need Me of the order of the weak scale [84], which
again would lead to a strong coupling regime at sufficiently large energy.
The present point of view is different, thanks to more recent develop-
ments that go under the name of generalized string dualities. The first
dualities to be explored were the so-called perturbative ones, which related
small and large compactification radii (for a review, see e.g. [85]). Sub-
sequently, new non-perturbative dualities were found, relating the strong
coupling limit of a given string theory to the weak coupling limit of another
string theory (for pedagogical introductions, see e.g. [18, 86]). The realiza-
tion that different superstring theories could be connected, in one way or
another, by some duality, led to the presently accepted picture, accord-
ing to which there is a single fundamental theory, the so-called M-theory,
underlying all five different ten-dimensional superstring theories and also
eleven-dimensional supergravity. These developments led to the possibil-
ity of having very different relations among Ms, Me and Mp than in the
perturbative heterotic string. A key new phenomenon in all this, already
anticipated in some pioneering field-theoretical works [87], but made more
plausible by explicit string-theoretical constructions [88, 89], was the pos-
sibility that some of the compactified extra dimensions may be 'invisible'
to all the SM fields. For example, the SM fields living in four-dimensional

space-time could be interpreted as localized on a three-brane (substantially,
a fixed domain wall for energies much smaller than the string scale) embed-
ded in a higher-dimensional bulk. On the contrary, the graviton (and pos-
sibly other exotic fields with interactions of gravitational strength) would
be free to propagate in the bulk space-time of higher-dimensionality. This
opened an entirely new game for string-inspired model-building, even if at
present we can already enumerate a large numbers of interesting attempts,
but we do not have yet a calculable and complete realistic picture.
As one example of the possible phenomenological implications of the
new approach, we describe here a moderate departure from the standard
superstring phenomenology, i.e. the scenario for the unification of couplings
emerging from the so-called Horava-Witten picture. A pictorial illustration
of the key point is given in Fig. (8), taken from [86]. Case a) shows the run-
ning of the three SM gauge couplings and of the dimensionless gravitational
coupling, Q:a '" E2 /M~, assuming the MSSM spectrum at the weak scale
and no new thresholds up to Mp, and is representative of heterotic string
models at weak coupling. Case b) shows the effect of a fifth dimension be-
low the unification scale: the Kaluza-Klein excitations of the MSSM states,
as well as of the graviton, lead to a drastically modified evolution and an
accelerated running of all the coupling constants above Me. Case c) shows
a possible novel phenomenon. In the strong coupling limit, the Es x Es
heterotic string leads to a new dimension which is slightly different from
the familiar ten dimensions that are usually considered in the perturbative
discussion of heterotic string compactifications. Instead of being similar to
a circle, it is more like a segment [88]. The gauge fields and matter live
at the endpoints only, while gravity propagates in the bulk. Suppose that
a fifth dimension of this type exists below the unification scale. Since the
MSSM fields live in the walls, the evolution of the gauge couplings is the
standard four-dimensional one. Since gravity propagates in the full five di-
mensions, however, the effective gravitational coupling runs faster than in
four dimensions. For a fifth dimension of the appropriate size, the kink in
the gravitational coupling can make all couplings meet [89] at the unifica-
tion scale Mu. Of course, this is not more predictive than ordinary grand
unification, since the size of the fifth dimension can be taken as a parame-
ter, but it shows that the string unification problem may be solved in some
appealing way.
Other, much more drastic possibilities, made possible by superstring
dualities, have recently come to fashion. One can push down the string
scale Ms as much as allowed by the present collider data, i.e. down to the
TeV scale [90]. One can also consider the possibility of very large large extra
dimensions, with typical radius close to 1 mm (the limit set by Cavendish-
322 FABIO ZWIRNER
10°
"3
~
10-2 ",
10-4
a)
10-6
10 12
GeV
10° 10°
"3 "3
~ ~
10-2 ", 10-2 ",
10-4 10-4
b) c)
10-6 10-6
10 12 10 '8 10 '2 10 '8

GeV GeV
Figure 8. a) Running of the three gauge couplings and the dimensionless gravitational
cou piing with energy; b) effect of a fifth dimension below the unification scale; c) effect
of a fifth dimension of the Horava-Witten type.
type experiments) [91], as long as only gravitationally interacting fields

propagate in these dimensions. As a concrete example, one can consider
anisotropic compactifications of type-I strings, with n 'large' dimensions of
size R and the remaining dimensions of 'small' size r '" Mil. Neglecting
some numerical and coupling constant factors, Gauss' law prescribes that
M~ '" M~+2 . Rn. The case n = 2 is particularly attractive, since it would
lead to Ms '" 1 TeV and R '" 1 mm. This, however, is just one of the
many possibilities: for the moment we have no dynamical principle that
makes is preferable, even if it is intriguing to identify the string scale with
the weak scale and the compactification scale with the scale of a possible
cosmological constant. Inspired by the above, one can study some 'model-
independent' phenomenology at colliders [92], for example the emission of
Kaluza-Klein gravitons, in association with a photon or a jet, at e+ e- or
hadron colliders, respectively. The dimensionless coupling of a Kaluza-Klein
graviton to a quark or a lepton scales as EIMp, so that the cross-section
for the production of a single very light Kaluza-Klein graviton scales as
(EIMp)2. If one considers an inclusive process, however, one must integrate
over the number of Kaluza-Klein gravitons accessible at the given energy,
which is roughly given by (ER)n '" EnM~/M~+2, where n 2 2 is the

number of extra dimensions. The total inclusive cross-section would then
go as (7 ' " aEn / M~+2, which may not be negligible at present or future
high-energy colliders if Ms is sufficiently low. Notice the similarities with
the superlight gravitino case. Experiments are already able to set interesting
limits, as discussed in other lectures at this School [3, 4].
The possibility of large extra dimensions is very exciting, but one should
be aware of the large number of new problems it brings in (without solving,
at least for the moment, any of the old problems of supersymmetric unified
theories). The successful prediction of supersymmetric grand-unification is
in general lost, even if it may be recovered in some special models. New
flavour problems arise, since the small compactification scale is now insuffi-
cient to suppress alone the dangerous higher-dimensional couplings, includ-
ing those associated with baryon- and lepton-number violation, and some
more powerful symmetry principle is needed. Nevertheless, exploring this
new possibility may eventually lead to some possible (and badly needed!)
insight on the gauge hierarchy, cosmological constant and supersymmetry-
breaking problems. This is the reason for the present intense theoretical
activity. Time will tell ...
8. Concluding remarks
The aim of these lectures was to explain, to an audience mainly composed

of young experimentalists, why low-energy supersymmetry is a motivated
and phenomenologically viable extension of the SM near the electroweak
scale, directly testable in the next few years.
Phenomenological studies of MSSM signals at present and future accel-
erators are at an advanced stage, and are continuously improving. Impor-
tant indirect tests of SUSY are also possible in the realm of flavour physics.
Given the present absence of definite experimental or theoretical evidence,
in setting up the framework for these searches we should not be prisoners
of too restrictive frameworks: Nature may have more imagination than we
do!
On the theoretical side, some major open problems remain: the dynam-
ics of SUSY breaking, the SUSY flavour puzzle, the cosmological constant
problem. Despite the intense theoretical activity on all of them, the feeling
is that some firm guiding principle is needed to make substantial progress.
The present hope is that string theories and their fascinating duality prop-
erties will provide it, when better understood. The subject is still young,
and there is a lot of room left for future investigations ...
324 FABIO ZWIRNER
Acknowledgements
The author would like to thank the organizers of the School for the hos-
pitality and for creating, together with the participants, a pleasant and
stimulating atmosphere.
References
1. See, e.g., the Minutes of the 53rd meeting of the LEP Experiments Committee,
November 1999 (http://www.cern.ch/Committees/LEPC /minutes/LEPC53.html).
2. The LEP Electroweak Working Group, results presented at the 1999 Summer Con-
ferences (http://www.cern.ch/LEPEWWGj)
3. E. Fernandez, these Proceedings, and references therein.
4. H. Montgomery, these Proceedings, and references therein.
5. C.S. Wood et al., Science 275 (1997) 1759; S.C. Bennett and C.E. Wieman, Phys.
Rev. Lett. 82 (1999) 2484.
6. G. Altarelli, R. Barbieri and F. Caravaglios, Int. J. Mod. Phys. A13 (1998) 1031;
A. Sirlin, hep-ph/9912227; J. Erler and P.Langacker, in update for the 2000 edi-
tion of PDG (http://pdg.lbl.gov/); J.A. Bagger, A.F. Falk and M. Swartz, hep-
ph/9908327.
7. M.E. Shaposhnikov, these Proceedings, and references therein.
8. C. Caso et ai, update for the 2000 edition of PDG (http://pdg.lbl.gov/).
9. For some modern reviews on effective field theories see, e.g.: J. Polchinski, hep-
th/9210046; H. Georgi, Ann. Rev. Nucl. Part. Sci. 43 (1994) 209; A. Manohar,
hep-ph/9606222.
10. L. Di Lelia, these Proceedings, and references therein.
11. P. Salati, these Proceedings, and references therein.
12. K. Wilson, as quoted by L. Susskind, Phys. Rev. D20 (1979) 2019; E. Gildener and
S. Weinberg, Phys. Rev. D13 (1976) 3333; S. Weinberg, Phys. Lett. B82 (1979)
387; G. 't Hooft, in Recent deVelopments in gauge theories (G. 't Hooft et al., eds.),
Plenum Press, 1980, p. 135.
13. N. Cabibbo, L. Maiani, G. Parisi and R. Petronzio, Nucl. Phys. B158 (1979) 295.
14. For a review and a list of original references see, e.g.: U.M. Heller, plenary talk at
LATTICE 93, Nucl. Phys. Proc. Suppl. B34 (1994) 101.
15. T. Hambye and K. Riesseimann, Phys. Rev. D55 (1997) 7255.
16. For some recent reviews and original references see, e.g.: K. Lane, hep-ph/9501249
and hep-ph/9610463; R.S. Chivukuia, hep-ph/970l322 and hep-ph/9803219.
17. R. Haag, J. Lopuszanski and M. Sohnius, Nucl. Phys. B88 (1975) 257.
18. E. Kiritsis, these Proceedings, and references therein.
19. S. Weinberg, Rev. Mod. Phys. 61 (1989) 1 and astro-ph/9610044; S.M. Carroll
and W.H. Press, Ann. Rev. Astronom. Astrophys. 30 (1992) 499; L.M. Krauss and
M.S. Turner, Gen. ReI. Grav. 27 (1995) 1137.
20. N. Bahcall, J. P. Ostriker, S. Perlmutter and P. J. Steinhardt, Science 284 (1999)
1481.
21. S. Ferrara, C. Kounnas and F. Zwirner, Nucl. Phys. B429 (1994) 589; N. Arkani-
Hamed, S. Dimopoulos and G. Dvali, Phys. Lett. B429 (1998) 263; R. Sundrum,
JHEP 9907 (1999) 001.
22. J. Wess and J. Bagger, Supersymmetry and Supergravity, 2nd edition (Princeton
University Press, 1992).
23. S.J. Gates, M.T. Grisaru, M. Rocek and W. Siegel, Superspace or One Thousand
and One Lessons in Supersymmetry (Benjamin-Cummings, 1983).
24. M.F. Sohnius, Phys. Rep. 128 (1985) 39.
25. J.-P. Derendinger, preprint ETH-TH/90-21.
26. P.C. West, Introduction to Supersymmetry and Supergravity (World Scientific,

1990).
27. H.E. Haber and G.L . Kane, Phys. Rep. 117 (1985) 75.
28. J. Wess and B. Zumino, Nucl. Phys. B70 (1974) 39 and B78 (1974) 1; S. Ferrara
and B. Zumino, Nuc!. Phys. B79 (1974) 413; A. Salam and J. Strathdee, Phys. Lett.
B51 (1974) 353 and Nuc!. Phys. B76 (1974) 477.
29. Yu.A. Gol'fand and E.P. Likhtman, JETP Lett. 13 (1971) 323; D.V. Volkov and
V.P. Akulov, Phys. Lett. B46 (1973) 109.
30. J. Wess and B. Zumino, Phys. Lett. B49 (1974) 52; J. Iliopoulos and B. Zumino,
Nucl . Phys. B76 (1974) 310; S. Ferrara, J. Iliopoulos and B. Zumino, Nucl. Phys.
B77 (1974) 413; B. Zumino, Nucl. Phys. B89 (1975) 535; P.C. West, Nucl. Phys.
B106 (1976) 219; M.T. Grisaru, W. Siegel and M. Rocek, Nucl. Phys. B159 (1979)
429.
31. L. Maiani, Proc. Summer School on Particle Physics, Gif-sur-Yvette, 1979 (IN2P3,
Paris, 1980), p. 1; M. Veltman, Acta Phys. Polon. B12 (1981) 437; E. Witten, Nuc!.
Phys. B188 (1981) 513.
32. S. Ferrara, L. Girardello and F. Palumbo, Phys. Rev. D20 (1979) 403.
33. L. Girardello and M.T. Grisaru, Nuc!. Phys. B194 (1982) 65.
34. H.-P. Nilles, Phys. Rep. 110 (1984) 1.
35. P. Fayet, Nucl. Phys. B90 (1975) 104, Phys. Lett. B64 (1976) 159 and B69 (1977)
489.
36. S. Dimopoulos and H. Georgi, Nuc!. Phys. B193 (1981) 150.
37. J. Ellis and D.V. Nanopoulos, Phys. Lett. Bll0 (1982) 44; R. Barbieri and R. Gatto,
Phys. Lett. B110 (1982) 211.
38. J . Ellis, S. Ferrara and D.V. Nanopoulos, Phys. Lett . B114 (1982) 231; W.
Buchmiiller and D . Wyler, Phys. Lett. B121 (1983) 321; J . Polchinski and M.B.
Wise, Phys. Lett. B125 (1983) 393; F. del Aguila, J.A. Grifols, A. Mendez, D.V.
Nanopoulos and M. Srednicki, Phys. Lett. B129 (1983) 77; J.-M. Frere and M.B.
Gavela, Phys. Lett. B132 (1983) 107.
39. M. Misiak, S. Pokorski and J. Rosiek, hep-ph/9703442; Y. Grossman, Y. Nir and
R. Rattazzi, hep-ph/9701231; A. Masiero and L. Silvestrini, hep-ph/9709242 and
hep-ph/9711401; D. A. Demir, A. Masiero and O. Vives, hep-ph/9911337.
40. T.K. Kuo and N. Nakagawa, Nuovo Cimento Lett. 36 (1983) 560; L. Alvarez-Gaume,
J. Polchinski and M.B. Wise, Nucl. Phys. B221 (1983) 495; R. Barbieri and L. Ma-
iani, Nuc!. Phys. B224 (1983) 32.
41. A. Djouadi et aI., Nuc!. Phys. B349 (1991) 48; M. Boulware and D. Finnell, Phys.
Rev. D44 (1991) 2054.
42. R. Barbieri, F. Caravaglios and M. Frigeni, Phys. Lett. B279 (1992) 169.
43. G. Cho and K. Hagiwara, hep-ph/9912260.
44. Y. Okada, M. Yamaguchi and T. Yanagida, Prog. Theor. Phys. Lett . 85 (1991) 1;
J . Ellis, G. Ridolfi and F . Zwirner, Phys. Lett . B257 (1991) 83 and B262 (1991)
477; H.E. Haber and R . Hempfling, Phys. Rev. Lett. 66 (1991) 1815; R. Barbieri,
M. Frigeni and M. Caravaglios, Phys. Lett. B258 (1991) 167.
45. M. Carena, H. E. Haber, S. Heinemeyer, W. Hollik, C. E. Wagner and G. Weiglein,
hep-ph/000l002.
46. P. Zerwas, these Proceedings, and references therein.
47. M. Carena et al., Report of the Higgs Working Group, to appear in the Proceedings
of the Fermilab Workshop on Supersymmetry /Higgs Physics at Run II.
48. Z. Kunszt and F. Zwirner, Nuc!. Phys. B385 (1992) 3.
49. M. Spira and P. M. Zerwas, hep-ph/9803257.
50. F.E. Paige, hep-ph/9909215, and references therein.
51. P. Langacker, Phys. Rep . 72 (1981) 185; G.G. Ross, Grand Unified Theories,
Benjamin-Cummings, 1984; R.N. Mohapatra, hep-ph/980l235 .
52. H. Georgi and S.L. Glashow, Phys. Rev. Lett. 32 (1974) 438.
53. J.C. Pati and A. Salam, Phys. Rev. D8 (1973) 1240.
326 FABIO ZWIRNER
54. M.S. Chanowitz, J. Ellis and M.K. Gaillard, Nucl. Phys. B128 (1977) 506;
A.J. Buras, J. Ellis, M.K. Gaillard and D.V. Nanopoulos, Nucl. Phys. B135 (1978)
66.
55. H. Georgi, H.R. Quinn and S. Weinberg, Phys. Rev. Lett. 33 (1974) 451.
56. U. Amaldi et al., Phys. Rev. D36 (1987) 1385; G. Costa et al., Nucl. Phys. B297
(1988) 244.
57. S. Dimopoulos and H. Georgi, Nucl. Phys. B193 (1981) 150; N. Sakai, Z. Phys. Cl1
(1982) 153.
58. S. Dimopoulos, S. Raby and F. Wilczek, Phys. Rev. D24 (1981) 1681; L.E. Ibanez
and G.G. Ross, Phys. Lett. B105 (1981) 439.
59. S. Weinberg, Phys. Rev. D26 (1982) 287; N. Sakai and T. Yanagida, Nucl. Phys.
B197 (1982) 533.
60. S. Weinberg, Phys. Lett. B91 (1980) 51; L.J. Hall, Nucl. Phys. B178 (1980) 75; T.J.
Goldman and D.A. Ross, Nucl. Phys. Bl71 (1980) 273; P. Binetruy and T . Schucker,
Nucl. Phys. B178 (1981) 293,307; I. Antoniadis, C. Kounnas and C. Roiesnel, Nucl.
Phys. B198 (1982) 317.
61. M.B. Einhorn and D.R.T. Jones, Nucl. Phys. B196 (1982) 475; M.E. Machacek and
M.T . Vaughn, Nucl. Phys. B236 (1984) 221.
62. R. Barbieri, S. Ferrara, 1. Maiani, F. Palumbo and C.A. Savoy, Phys. Lett. B115
(1982) 212.
63. K. Inoue, A. Kakuto, H. Komatsu and S. Takeshita, Progr. Theor. Phys. 67 (1982)
1889; L. Alvarez-Gaume, J. Polchinski and M.B. Wise, Nucl. Phys. B221 (1983)
495; J. Bagger, S. Dimopoulos and E. Mass6, Phys. Rev. Lett. 55 (1985) 920.
64. N. Cabibbo, L. Maiani, G. Parisi and R. Petronzio, Nucl. Phys. B158 (1979) 295;
B. Pendleton and G.G. Ross, Phys. Lett. B98 (1981) 291; C. Hill, Phys. Rev D24
(1981) 691.
65. A. Donini, Nucl.Phys. B467 (1996) 3; D.M. Pierce, J .A. Bagger, K. Matchev and
R. Zhang, Nucl. Phys. B491 (1997) 3.
66. M. Carena and C.E.M. Wagner, hep-ph/9407208; G.K. Leontaris and N.D. Tracas,
Phys. Lett. B351 (1995) 487; C. Kounnas, I. Pavel, G. Ridolfi and F. Zwirner, Phys.
Lett. B354 (1995) 322.
67. K. Inoue, A. Kakuto, H. Komatsu and S. Takeshita, Progr. Theor. Phys. 67 (1982)
1889, 68 (1982) 927 and 71 (1984) 413; L. Alvarez-Gaume, J. Polchinski and
M.B . Wise, Nucl . Phys. B221 (1983) 495; J .-P. Derendinger and C.A. Savoy, Nucl.
Phys. B237 (1984) 307; B. Gato, J. Leon, J. Perez-Mercader and M. Quiros, Nucl.
Phys. B253 (1985) 285; N.K. Falck, Z. Phys. C30 (1986) 247.
68. L.E. Ibanez and G.G. Ross, Phys. Lett. 110B (1982) 215; K. Inoue, A. Kakuto, H.
Komatsu and S. Takeshita, Progr. Theor. Phys. 68 (1982) 927 and 71 (1984) 413;
L. Alvarez-Gaume, M. Claudson and M.B. Wise, Nucl. Phys. B207 (1982) 96; J.
Ellis, D.V. Nanopoulos and K. Tamvakis, Phys. Lett. B121 (1983) 123.
69. J.A. Casas, A. Lleyda and C. Munoz, Nucl. Phys. B471 (1996) 3 and Phys. Lett.
B389 (1996) 305; A. Strumia, Nucl. Phys. B482 (1996) 24; H. Baer, M. Brhlik and
D. Castano, Phys. Rev. D54 (1996) 6944; J.A. Casas, hep-ph/9707475.
70. P. Fayet and J. Iliopoulos, Phys. Lett. B51 (1974) 461; L. O'Raifeartaigh, Nucl.
Phys. B96 (1975) 331; P.Fayet, Phys. Lett. B58 (1975) 67.
71. S. Deser and B. Zumino, Phys. Rev. Lett. 38 (1977) 1433; E. Cremmer, B. Julia,
J. Scherk, S. Ferrara, L. Girardello and P. van Nieuwenhuizen, Nucl. Phys. B147
(1979) 105.
72. E. Cremmer, S. Ferrara, L. Girardello and A. Van Proeyen, Nucl. Phys. B212 (1983)
413; J. Bagger, Nucl. Phys. B211 (1983) 302.
73. T.R. Taylor, hep-ph/9510281.
74. S. Ferrara, L. Girardello and H.P. Nilles, Phys. Lett. B125 (1983) 457; J.-
P. Derendinger, L.E. Ibanez and H.P. Nilles, Phys. Lett. B155 (1985) 65; M. Dine,
R. Rohm, N. Seiberg and E. Witten, Phys. Lett. B156 (1985) 55; C. Kounnas and
M. Porrati, Phys. Lett. B191 (1987) 91.
75. R. Rohm, Nucl. Phys. B237 (1984) 553; C. Kounnas and M. Porrati, Nue!. Phys.
B310 (1988) 355; S. Ferrara, C. Kounnas, M. Porrati and F. Zwirner, Nucl. Phys.
B318 (1989) 75; M. Porrati and F. Zwirner, Nucl. Phys. B326 (1989) 162; C. Koun-
nas and B. Rostand, Nucl. Phys. B341 (1990) 641; I. Antoniadis, Phys. Lett. B246
(1990) 377; I. Antoniadis and C. Kounnas, Phys. Lett. B261 (1991) 369; I. An-
toniadis, C. Munoz and M. Quiros, Nue!. Phys. B397 (1993) 515; I. Antoniadis,
E. Dudas and A. Sagnotti, Nue!. Phys. B544 (1999) 469 and Phys. Lett. B464
(1999) 38; C. Angelantonj, I. Antoniadis, G. D'Appollonio, E. Dudas and A. Sag-
notti, hep-th/991108l.
76. A.E. Nelson, hep-ph/9707442; E. Poppitz, hep-ph/9710274; S. Thomas, hep-
th/9801007.
77. A.H. Chamseddine, R. Arnowitt and P. Nath, Phys. Rev. Lett. 49 (1982) 970;
R. Barbieri, S. Ferrara and C.A. Savoy, Phys. Lett. BU9 (1982) 343.
78. P. Fayet, Phys. Lett. B175 (1986) 471, and references therein.
79. G.F. Giudice and R. Rattazzi, hep-ph/980127l.
80. A. Brignole, F. Feruglio and F. Zwirner, Nue!. Phys. B501 (1997) 332 and JHEP
11 (1997) 001.
81. N. Cabibbo, G.R. Farrar and L. Maiani, Phys. Lett. BI05 (1981) 155; P. Fayet,
Phys. Lett. B175 (1986) 471.
82. A. Brignole, F. Feruglio and F. Zwirner, hep-ph/9711516; A. Brignole, F. FerugIio,
M.L. Mangano and F. Zwirner, hep-ph/9801329.
83. V.S. Kaplunovsky, Nue!. Phys. B307 (1988) 145; L. Dixon, V.S. Kaplunovsky and
J. Louis, Nue!. Phys. B355 (1991) 649; J.P. Derendinger, S. Ferrara, C. Kounnas
and F. Zwirner, Nue!. Phys. B372 (1992) 145 and Phys. Lett. B271 (1991) 307; G.
Lopez-Cardoso and B.A. Ovrut, Nue!. Phys. B369 (1992) 351; I. Antoniadis, K.S.
Narain and T. Taylor, Phys. Lett. B267 (1991) 37; I. Antoniadis, J. Ellis, R. Lacaze
and D.V. Nanopoulos, Phys. Lett. B268 (1991) 188; L.E. Ibanez, D. Lust and G.G.
Ross, Phys. Lett. B272 (1991) 251.
84. R. Rohm and E. Witten, Annals Phys. 170 (1986) 454; I. Antoniadis, C. Bachas,
D. C. Lewellen and T. N. Tomaras, Phys. Lett. B207 (1988) 441 (1988); S. Ferrara,
C. Kounnas, M. Porrati and F. Zwirner, Nucl. Phys. B318 (1989) 75.
85. A. Giveon, M. Porrati and E. Rabinovici, Phys. Rep. 244 (1994) 77.
86. J. Polchinski, Rev. Mod. Phys. 68 (1996) 1245.
87. V. A. Rubakov and M. E. Shaposhnikov, Phys. Lett. B125 (1983) 136.
88. P. Horava and E. Witten, Nucl . Phys. B460 (1996) 506.
89. E. Witten, Nue!. Phys. B471 (1996) 135.
90. J. D. Lykken, Phys. Rev. D54 (1996) 3693.
91. N. Arkani-Hamed, S. Dimopoulos and G. Dvaii, Phys. Lett. B429 (1998) 263.
92. G. F. Giudice, R. Rattazzi and J. D. Wells, Nucl. Phys. B544 (1999) 3;
E. A. Mirabelli, M. Perelstein and M. E. Peskin, Phys. Rev. Lett. 82 (1999) 2236.
ENHANCED GLOBAL SYMMETRIES AND A
STRONG ELECTROWEAK SECTOR
ChiraljConformal Phase Transition
F. SANNINO
Yale University
New Haven CT 06520-8120, USA
Abstract. In this talk I review the intriguing possibility of Ref. [1] that the
physical spectrum of a vector-like gauge field theory exhibits an enhanced
global symmetry near a chiral phase transition. A transition from the Gold-
stone phase to the symmetric phase is expected as the number of fermions
N f is increased to some critical value. Various investigations have suggested
that a parity-doubled spectrum develops as the critical value is approached.
Using an effective Lagrangian as a guide one observes that a parity doubling
is associated with the appearance of an enhanced global symmetry in the
spectrum of the theory. If such a near-critical theory describes symmetry
breaking in the electroweak theory, the additional symmetry suppresses the
contribution of the parity doubled sector to the S parameter.
1. Introduction
Gauge field theories exhibit many different patterns of infrared behavior. In-
deed for a vector-like theory such as QeD, it is known that for low values of
Nf' the theory confines and chiral symmetry breaking occurs. On the other
hand, for large N f the theory loses asymptotic freedom. In between, there
is a conformal window where the theory does not confine, chiral symmetry
is restored, and the theory acquires a long range conformal symmetry. It
has been proposed that for an SU (N) gauge theory, there is a transition
from the confining, chirally broken theory to the chirally symmetric theory
at Nf :::::: 4N [2, 3]. Recent lattice simulations, however, seem to indicate [4]
that the amount of chiral symmetry breaking decreases substantially (for
N = 3) when N f is only about 4.
329
330 F. SANNINO
Assuming that a single transition takes place at some critical value

of Nf, we can ask questions about the spectrum of the theory near the
transition. In Reference [5], it was argued by studying Weinberg spectral
function sum rules that for near-critical theories parity partners become
more degenerate than in QCD-like theories. This leads naturally to the
idea that parity doublets might form as chiral symmetry is being restored.
Lattice studies also indicate such a possibility [4].
Here I review the ideas presented in Ref. [1] where it was observed
using an effective-Lagrangian as a guide, that the formation of degenerate
parity partners is associated with the appearance of an enhanced global
symmetry in the spectrum of states. This new symmetry could play a key
role in describing a possible strong electroweak Higgs sector. Whether the
new symmetry can be shown to emerge dynamically from an underlying
gauge theory with Nf near a critical value remains an open question.
First, in Section 2, the appearance of enhanced global symmetry is dis-
cussed. Confinement is assumed and the symmetry of the underlying gauge
theory, SUL(Nf) x SUR(Nf), is built into an effective Lagrangian describing
the physical states of the theory. Parity invariance is imposed and the usual
pattern of chiral symmetry breaking (SUL(Nf ) x SUR(Nf) ~ SUv(Nf)) is
assumed. The N} - 1 Goldstone bosons appear together with scalar chiral
partners. We augment the spectrum with a set of vector fields for both the
SUL (Nf) and SUR (Nf) symmetry groups. The Lagrangian thus takes the
form of a linear sigma model coupled to vectors.
Analyzing the spectrum one recognizes that there is a particular choice
of the parameters that allows for a degenerate vector and axial-vector,
while enlarging the global symmetry to include an additional (unbroken)
SUL(Nf) x SUR(Nf). This happens as the spectrum of the theory splits
into two sectors with one displaying the additional symmetry. I then review
the arguments (see Ref. [5]) that an underlying near-critical SU(N) gauge
theory might naturally lead to a more degenerate vector-axial spectrum
than in QCD, and to an enhanced symmetry. Finally we note that even a
discrete additional symmetry, Z2L x Z2R, of the effective theory is adequate
to insure the mass degeneracy of the vector and axial vector.
The possible appearance of an additional, continuous symmetry was
considered by Casalbuoni et al in Refs.[6, 7].
Finally in Section 3 the electroweak gauge group is embedded within
the global symmetry group. As observed in Ref. [1] the enhanced symmetry
of the strongly interacting sector, which now provides electroweak symme-
try breaking, plays an important role. The additional symmetry operates
as a partial custodial symmetry for the electroweak S parameter, in the
sense that the parity doubled part of the strong sector, by itself, makes no
contribution to S.
ENHANCED GLOBAL SYMMETRIES 331
2. Effective Lagrangian for SUL(Nf ) x SUR(Nf) global symmetry

To discuss the possible appearance of enhanced symmetry some descrip-
tion of the spectrum is needed. It is helpful to use an effective Lagrangian
possessing SUL(Nf) x SUR(Nf) symmetry, the global invariance of the un-
derlying gauge theory. Chiral symmetry is broken according to the standard
pattern SUL(Nf) x SUR(Nf) -+ SUv(Nf). The Nj- 1 Goldstone bosons
are encoded in the Nf x Nf real traceless matrix <P~ with i,j = 1, ... , Nf.
The complex matrix M = S + i<p describes both the Goldstone bosons as
well as associated scalar partners S. It transforms linearly under a chiral
rotation of the type M -+ uLMuk with UL /R in SUL/R(Nf).
To augment the massive spectrum one introduces vector and axial vector
fields following a method outlined in Ref. [9]. One first formally gauge the
global chiral group introducing the covariant derivative
DtJ.M = atJ. M - if;A'{M + if;MA~ , (1)
where A't/R = A'tiaR Ta and T a are the generators of SU(Nf), with a =

1, ... , Nl- 1 and Tr [TaT b] = ~Jab. Under a chiral transformation
A tJ.L/R -- UL/R AtJ.LUL/R

t 2 atJ. t
- -;; - UL/RU L / R· (2)
9
The effective Lagrangian needs only to be invariant under global chiral
transformations. Including terms only up to mass dimension four, it may
be written in the form
L ~Tr [DtJ.MDJ.'Mt] + m 2 Tr[ALJ.'A'{ + ARJ.'A~]

+ h Tr [ALJ.'M A~Mt] + r Tr [ALJ.'A'{M Mt + ARJ.'A~Mt M]
+ i ~ Tr [ALtJ. (M DJ.' Mt - DtJ. M Mt)
+ ARJ.' (MtDJ.'M - DJ.'MtM)] . (3)
The parameters h, r and s are dimensionless real parameters, while m 2 is

a common mass term. To this, we may add a kinetic term for the vector
fields
(4)
where Ff/R = aJ.' AL/R - a v A't/R - if; [A't/R' A L / R] along with vector-
interaction terms respecting only the global symmetry. Finally, one may
add the double trace term, Tr [MMt] Tr [Ai + A~] at the dimension-four
332 F. SANNINO
level. To arrange for symmetry breaking, a potential V (M, Mt) must be

added. When the effective Lagrangian is extended to the dimension-six
level and higher, many new operators enter. Parity is also a symmetry. It
is worth noting that a Lagrangian of this type has been used to describe
the low-lying QCD resonances and interactions [10].
The scalar vacuum expectation value is v and the new vector fields are
(5)
The vector-axial vector mass difference is given by
(6)
In QCD this difference is known experimentally to be positive, a fact that

can be understood by examining the Weinberg spectral function sum rules
(see Ref. [11] and references therein). The effective Lagrangian description
is of course unrestrictive. Depending on the values of the g, sand h param-
eters, one can have a degenerate or even inverted mass spectrum.
What kind of underlying gauge theory might provide a degenerate or in-
verted spectrum? Clearly, it has to be different from QCD, allowing for
a modification of the spectral function sum rules. In Reference [5], an
SU(N) gauge theory (with N > 2) and Nf flavors was considered. If Nf
is large enough but below 11N/2, an infrared fixed point of the gauge
coupling u* exists, determined by the first two terms in the f3 function.
For Nf near 11N/2, u* is small and the global symmetry group remains
unbroken. For small N f' on the other hand, the chiral symmetry group
SUdNf) x SUR(N,} breaks to its diagonal subgroup. One possibility is
that the transition out of the broken phase takes place at a relatively
large value of NJlN(-;::;:, 4), corresponding to a relatively weak infrared fixed
point [2, 3]. An alternate possibility is that the transition takes place in
the strong coupling regime, corresponding to a small value of NJiN [4].
The larger value emerges from the renormalization group improved gap
equation, as well as from instanton effects [12], and saturates a recently
conjectured upper limit [13]. It corresponds to the perturbative infrared
fixed point u* reaching a certain critical value U c . A similar result has also
been obtained by using a suitable effective Lagrangian [3].
In Reference [5] the spectrum of states in the broken phase near a large-
NJiN transition was investigated using the spectral function sum rules. It
was shown that the ordering pattern for vector-axial hadronic states need
not be the same as in QCD-like theories (small NJlN). The crucial ingredi-
ent is that these theories contain an extended "conformal region" extending
from roughly 21r F'/r to the scale A where asymptotic freedom sets in. In this
region, the coupling remains close to an approximate infrared fixed point

and the theory has an approximate long range conformal symmetry. It was
argued that this leads to a reduced vector-axial mass splitting, compared
to QeD-like theories. This suggests the interesting possibility that parity
doublets begin to form as chiral symmetry is being restored. That is, the
vector-axial mass ratio approaches unity as the masses decrease relative to
A. Lattice results seem to provide supporting evidence for such a possibil-
ity [4], although at smaller values of Nt/N.
If a parity doubled spectrum does appear, it is natural to expect it to
be associated with some new global symmetry. While it is hard to demon-
strated the appearance of a new global symmetry using the underlying de-
grees of freedom, one can explore aspects of parity doubling at the effective
Lagrangian level. Returning to this description, we note that vector-axial
parity doubling corresponds to the parameter choice (see Eq. (6)),
r/ = 2g s + h . (7)
This condition does not yet reveal an additional symmetry and therefore
there is no reason to expect parity degeneracy to be stable in the presence
of quantum corrections and the many higher dimensional operators that
can be added to the effective Lagrangian in Eq. (3).
However, for the special choice s = g, r = g2/2 and h = _g2, the effec-
tive Lagrangian acquires a new continuous global symmetry that protects
the vector-axial mass difference. The effective Lagrangian at the dimension-
four level takes the simple form
L= ~ Tr [8J.LM8J.L Mt] + m 2 Tr [ALJ.LAi + ARJ.LA~], (8)

along with vector kinetic and interaction terms, the interaction term Eq. (2),
and the symmetry breaking potential V(M, Mt). The theory now has two
sectors, with the vector and axial vector having their own unbroken global
SUL(Nf) x SUR(Nf) . The two sectors interact only through the product
of singlet operators. The full global symmetry is [SUL(Nf) x SUR(Nf)]2 x
Uv(l) spontaneously broken to SUv(Nf) x Uv(l) x [SUL(Nf) x SUR(Nf )].
The vector and axial vector become stable due to the emergence of a new
conservation law. This enhanced symmetry would become exact only in the
chirallimit. For finite but small (relative to A) values of the mass scales in
Eq. (8), there are additional, smaller terms giving smaller mass splittings
and small width-to-mass ratios.
It is of course a simple observation that a new symmetry and conser-
vation law emerge if a theory is split into two sectors by setting certain
combinations of parameters to zero. But here we were led to this possibil-
ity by looking for a symmetry basis for the parity doubling that has been
hinted at by analyses of the underlying gauge theory.
334 F. SANNINO
It is worth mentioning that along with the additional global symmetry

SUL(N,) x SUR(N,), the effective Lagrangian Eq. (8) possesses a discrete
Z2L x Z2R symmetry. Under Z2L X Z2R the vector fields transform according
to AL -+ zLAL, AR -+ zRAR with zL/R = 1, -1 and ZL/R E Z2L/R·
Actually, the discrete symmetry alone is enough to insure vector-axial mass
degeneracy and stability against decay. In that case, additional interaction
terms, such as the single trace term rTr [A"'LA~M Mt + A"'RA~Mt M] are
allowed, but degeneracy and stability are still insured.
3. Strongly Interacting Electroweak Sector

To discuss the consequences of enhanced symmetry for a strong symme-
try breaking sector of the standard electroweak theory one embeds the
SUL(2) x Uy(l) gauge symmetry in the global SUL(N,) x SUR(N,) chiral
group. For simplicity I restrict attention to the SUL(2) x SUR(2) subgroup
of the full global group. The electroweak gauge transformation then takes
the form M -+ uwMu~. M is now a 2 x 2 matrix which can be writ-
ten as M = ~ [0" + if ·11] , where Uw = UL = exp O€ara)
with the ra
Pauli matrices, and where Uy = exp (~€or3). The weak vector boson fields
transform as
where 9 and g' are the standard electroweak coupling constants, W'" =
W '"ara dB B r3
Tan", = ",T·
A convenient method of coupling the electroweak gauge fields to the
globally invariant effective Lagrangian in Section 2 is to introduce a co-
variant derivative, which includes the Wand B fields as well as the strong
vector and axial-vector fields,
D'" M = 8'" M - ig W'" M + ig' M B'" - igcCfM + igc' M C~ , (10)
where we have defined the new vector fields

,
Cf = A~ - gW"', C~
9
= A~ -
f B'" an~ ~he~e c a~d c' are arbitrary re.al constan:s (for details see Ref. [1D.
tly reqUIrIng mvanance under the panty operatIon exchanging the labels
L f-t R we have the extra condition c = c'.
The effective Lagrangian, constructed to be invariant under a local
SUL(2) x Uy(1) as well as CP is through dimension four,
L = ~Tr [D",MD'" Mt] + m 2 Tr [CL",Cf + CR",C~] + h Tr [CL",MC~Mt]

+ rTr [CLJ'CfMMt + CRJ'C~MtM]

+ i ~ Tr [CLJ' (MDJ' Mt - DJ' M Mt) + CRJ' (MtDJ' M - DJ' Mt M) ]
(11)
To this we add a kinetic term
LKin =- ~Tr [FLJ'IIFfll + FRJ'IIF~II] - ~Tr [WJ'II WJ'II] - ~Tr [BJ'IIBJ'II] ,

(12)
where WJ'II = aJ' W II - allwJ' - ig [WI" WII] and BJ'II = aJ'BII - allBw One
should still add other interaction terms involving the C L / R fields, the inter-
action term Tr [MMt] Tr [cI + cit] and a symmetry breaking potential.
The extension of this effective Lagrangian to the relevant case of the
larger symmetry group SUdNf) x SUR(Nf ) with Nf > 2, is straightfor-
ward.
The vector and axial vector masses computed in Ref. [1], M~ and M1,
are arbitrary, depending on the choice of parameters, although generically
we expect them and the scalar masses to be of order 47r 2 V 2 •
Expanding the Lagrangian up to quadratic terms in the fields one ob-
serves the presence of weak mixing terms providing a contribution from
physics beyond the standard model to the oblique electroweak corrections.
(One can find the actual computations and more explicit formulae in Ref. [1]) .
These may be described by the S, T, and U parameters, but the last two
vanish in the present model because there is no breaking of weak isospin
in the strong sector. The S parameter receives contributions from all the
physics beyond the standard model, including, in the model being used here,
loops of pseudo-Goldstone bosons (PGB's), the strongly interacting mas-
sive scalars, and the vector and axial vector. The direct, vector-dominance
contribution of the vector and axial vector is
Svect-dom = -=2
87r [M1 (1 - X)2
M2 _ M2
M~
- M2 _ M2
1~ -=287r [1 - (1 - X)
2] . (13)
9 Z A Z V 9
2
with X = 2~2 9 [gc - s]. Clearly, this contribution to the S parameter can
A
take on any value depending on the choice of parameters. Its typical order
of magnitude, with the strong coupling estimate g2 ~ 47r 2 , is expected to
be 0(1).
The parameter choice s = gc and h = -g22 gives X = 0, leads to the
degeneracy of the vector and axial vector and the vanishing of Svect-dom.
g2c2
The further choice r = -2- leads to the collapse of the general effective
336 F. SANNINO
Lagrangian into the simple form
L = ~Tr [DJ.lMDI'Mt] + m 2 Tr [CLI'Ct + CRI'C~] , (14)

along with the kinetic terms ofEq. (12), interactions among the Cf/R fields
and a symmetry breaking potential. Here, D M = 8M - ig W M + ig' M B
is the standard electroweak covariant derivative.
The strongly interacting sector has split into two subsectors, communi-
cating only through the electroweak interactions. One subsector consists of
the Goldstone bosons together with their massive scalar partners. The other
consists of the degenerate vector and axial vector described by the Ai,/R
fields. In the absence of electroweak interactions, there is an enhanced sym-
metry [SUL(2) x SUR(2)] x [SUL(2) x SUR(2)], breaking spontaneously to
SUv(2) x [SUL(2) x SUR(2)]. The electroweak interactions explicitly break
the enhanced symmetry to SUL(2) x Uy(l). All of this may be generalized
to Nt > 2, necessary to yield a near-critical theory.
The additional symmetry has an important effect on the S parameter,
suppressing contributions that are typically large in QeD-like theories.
Finally, it could be that only a lesser, discrete symmetry emerges in
the physical spectrum. Even this would be sufficient to insure vector-axial
degeneracy and the vanishing of the vector dominance contribution to the
S parameter.
4. Conclusions
To explore some features that might arise in a strongly coupled gauge theory
when the number of fermions Nt is near a critical value for the transition
to chiral symmetry an effective Lagrangian approach was used.
The spectrum was taken to consist of a set of Goldstone particles, as-
sociated massive scalars, and a set of massive vectors and axial vectors. It
was observed that parity doubling is associated with the appearance of an
enhanced global symmetry, consisting of the spontaneously broken chiral
symmetry of the underlying theory (SUL(Nt) x SUR(Nt)) together with
an additional, unbroken symmetry, either continuous or discrete. The ad-
ditional symmetry leads to the degeneracy of the vector and axial vector,
and to their stability with respect to decay into the Goldstone bosons.
Despite the hints in Refs. [5, 4] it has not been established that an under-
lying gauge theory leads to these enhanced symmetries as Nt approaches
a critical value for the chiral transition. If it is to happen an interesting
interplay between confinement and chiral symmetry breaking would have
to develop at the transition.
Finally by electroweak gauging a subgroup of the chiral flavor group
it was shown that the enhanced symmetry provides a partial custodial
symmetry for the S parameter, in that there is no contribution from the

parity-doubled sector by itself.
Acknowledgments
I am happy to thank Thomas Appelquist and Paulo Sergio Rodrigues Da
Silva for sharing the work on which this talk is based. The work of F .S. has
been partially supported by the US DOE under contract DE-FG-02-92ER-
40704.
References
1. T. Appelquist, P.S. Rodrigues da Silva and F. Sannino, Report Numbers: YCTP-
P14-99, hep-ph/9906555. To be published on Phys. Rev. D.
2. T . Appelquist, J . Terning and L.C.R. Wijewardhana, Phys. Rev. Lett. 77, 1214
(1996).
3. F. Sannino and J. Schechter, Phys. Rev. D60, 056004, (1999).
4. R.D. Mawhinney, Nucl. Phys. Proc. Suppl. A 60, 306 (1998), Report Numbers: CU-
TP-802, hep-lat/9705030; D. Chen, R .D. Mawhinney, Nucl. Phys. Proc. Suppl. 53,
216 (1997) .
5. T. Appelquist and F. Sannino, Phys. Rev. D59, 067702 (1999).
6. R. Casalbuoni, S. De Curtis, D. Dominici, F. Feruglio and R. Gatto, Int. J. Mod.
Phys. A4, 1065 (1989); R. Casalbuoni, A. Deandrea, S. De Curtis, D. Dominici,
F. Feruglio, R . Gatto and M. Grazzini, Phys. Lett. B349, 533 (1995); R. Casalbuoni,
A. Deandrea, S. De Curtis, D. Dominici, R. Gatto and M. Grazzini, Phys. Rev . D53,
5201 (1996).
7. R. Casalbuoni, S. De Curtis, D. Dominici and M. Grazzini, Phys. Lett. B388 , 112
(1996); Phys. Rev D56, 5731 (1997) .
8. M. Bando, T. Kugo and K. Yamawaki, Phys. Rept. 164, 217 (1988).
9. O. Kaymakcalan and J . Schechter, Phys. Rev. D31, 1109 (1985) .
10. F. Sannino and J . Schechter, Phys. Rev. D52, 96 (1995); M. Harada, F. Sannino
and J. Schechter, Phys. Rev . D54, 1991 (1996); Phys. Rev. Lett. 78, 1603 (1997);
D. Black, A.H. Fariborz, F. Sannino and J. Schechter, Phys. Rev . D58, 54012 (1998) .
11 . C. Bernard, A. Duncan, J. LoSecco and S. Weinberg, Phys. Rev. D12, 792 (1975).
12. T . Appelquist and S. Selipsky, Phys. Lett. B400, 364 (1997); M. Velkovsky and E.
Shuryak, Phys. Lett. B437, 398 (1998).
13. T . Appelquist, A. Cohen and M. Shmaltz, Report Numbers: Phys. Rev. D60, 045003
(1999).
14. M. E. Peskin and T. Takeuchi, Phys. Rev. Lett. 65, 964 (1990); Phys. Rev. D46,
381 (1992) .
15. A. Longhitano, Phys. Rev. D22, 1166 (1980) ; Nucl. Phys. B188, 118 (1981).
16. T . Appelquist and J. Terning, Phys. Lett. B315, 139 (1993); S. Fleming and
I. Maksymyk, Phys. Rev. D53, 2781 (1996).
17. T . Appelquist, J. Terning, and L.C .W. Wijewardhana, Phys. Rev. Lett. 79, 2767
(1997).
18. S. Hands, J . Kogut, M. Lombardo and S. Morrison, Report Number: hep-
lat/9902034.
19. M. E. Peskin, Nucl. Phys. B175, 197 (1980), J . Preskill, Nucl. Phys. B177, 21
(1981).
SUPERSYMMETRY AND DUALITY IN FIELD THEORY
AND STRING THEORY
ELIAS KIRlTSIS
Physics Department, University of Crete
GR-71003 Heraklion, Greece
e-mail: kiritsis~physics.uch.gr
1. Introduction
A very important problem in physics is understanding strong coupling phe-

nomena. In the realm of high energy physics an appropriate example is the
low-energy regime of quantum chromo dynamics. Such examples appear also
very frequently in condensed matter systems.
There have been many attempts and methods to attack strong coupling
problems. These range from qualitative methods, to alternative approxima-
tions (non-standard perturbative expansions), to simple truncations of an
exact equation (typically applied to Schwinger-Dyson equations or renor-
malisation group equations), or finally direct numerical methods (usually
on a lattice).
All methods listed above have their merits, and can be suitable for the
appropriate problem. They also have their limitations. For example , de-
spite the successes of the lattice approach, some questions about QCD still
remain today beyond the reach of quantitative approaches. A typical exam-
ple are dynamic properties like scattering amplitudes. Consequently, new
analytical methods to treat strong coupling problems are always welcome.
The purpose of these lectures was to communicate to an audience of
mostly young experimentalists and standard model theorists, the progress
in this domain during the past few years.
The recent understanding of the strongly coupled supersymmetric field
theories is the starting point of the exposition as well as it central element,
electric-magnetic duality. We will go through the Seiberg-Witten solution
for N=2 gauge theories and we will briefly browse on other developments
of these techniques.
339
J.-J. Aubert et al. (eds.), Particle PhYSics: Ideas and Recent Developments, 339-395 .
340 ELIAS KIRITSIS
The most spectacular impact of these duality ideas has been in string
theory, a candidate theory for unifying all interactions including gravity. In
string theory, duality has unified the description and scope of distinct string
theories. The importance of new non-perturbative states was realized, and
their role in non-perturbative connections was elucidated. New advances
included the first microscopic derivation of the Bekenstein entropy formula
for black holes. Moreover, a new link was discovered relating gauge theories
to gravity, providing candidates for gauge theory effective strings. It is fair
to say that we have just glimpsed on new structures and connection in
the context of the string description of fundamental interactions. Whether
nature shares this point of view remains to be seen.
There are many excellent reviews that cover some of the topics I present
here and the readers are urgent to complement their reading by referring
to them. I will try to present a short representative list that will be the
initial point for those interested to explore the literature. There are several
reviews on supersymmetric field theory dualities [1]-[9]. Introductory books
and lectures in string theory can be found in [10]-[16]. Lectures on recent
advances and various aspects ofnon-perturbative string theory can be found
in [17]-[30].
2. Electric-Magnetic duality in Maxwell theory

We will describe in this section the simplest realization of a duality symme-
try, namely electric-magnetic duality in electrodynamics. We will be em-
ploying high energy units n = c = 1. The conventional Maxwell equations
are
~ -+ 8E -+
"VxB--=J (2.1)
at
(2.2)
We can use relativistic notation and assemble the electric and magnetic
fields into a second rank antisymmetric tensor FIJ.v as
Fij = -€ijkB k , jIJ. = (p, J)- (2.3)
If we define the dual electromagnetic field tensor as
(2.4)
Then Maxwell's equations (2.1),(2.2) can be written as
(2.5)
SUPERSYMMETRY AND DUALITY 341
The first of these is a true dynamical equation that we will continue to call
the Maxwell equation while the second becomes an identity once the fields
are written in terms of the electromagnetic potentials, FJ.'v = oJ.'Av - OvAl-'"
It is called the Bianchi identity.
Let us first consider the vacuum equations: p = 0, J = O. They can be
written as
- - - - - -
\1 . (E + iB) = 0 , i\1 x (E + iB)
+ iB) = 0
+ 8(E at (2.6)
which makes manifest the following symmetry of the equations

E + ill -t eirP(E + is) (2.7)
It turns out that only a discrete Z2 subgroup of this U(I) symmetry (¢ =
7r/2) has a chance of surviving the inclusion of charged matter. This is
known as the electric-magnetic duality transformation
, B-t -E (2.8)
or in tensor form
FJ.lv t-7 FJ.lv (2.9)
Once we consider the addition of charges, this symmetry can be main-
tained only at the expense of introducing also magnetic monopoles.
The classical (relativistic) equation of motion of a charged particle (with
charge e) in the presence of an electromagnetic field FJ.'v is given by
(2.10)
A magnetic monopole couples to F in the same way that a charge couples
to F. Classically, the generalisation of the equation above for a particle car-
rying both an electric charge e and a magnetic charge 9 is a generalisation
of (2.10)
(2.11)
Classically there are no conceptual changes apart from the fact that the
equation of motion is modified. The reason is that physics classically de-
pends on the field strengths rather than gauge potentials.
The situation changes in the quantum theory as was first pointed out
by Dirac. Physics does depend on the potentials rather than field strengths
alone, and this provides the famous Dirac quantisation condition for the
magnetic charge.
An easy way to see this is to write first the classical equation of motion
of a charged particle in the magnetic field ii of a magnetic monopole.
•• •
-
mr=erx B , B r
=9- -
-#
(2.12)
47r r3
342 ELIAS KIRITSIS
We can compute the (semi-classical) rate of change of the orbital angular

momentum
.)
dE d(
dt = dt m (r' x f) = m r x r.. (2.13)
Using the equation of motion we can substitute r and find
dE = eg r( f' x f) = !!--. (e g ~ (2.14)

dt 411" r3 dt 411" ~)
This indicates that the conserved angular momentum is given by
- _. eg r
Loot = m(r x f) - - - (2.15)
411" r
It can be verified that the second piece is the angular momentum of the
electromagnetic field, namely proportional to the spatial integral of the
Poynting vector E x E. Quantisation of the total and orbital angular mo-
mentum translates via (2.15) to the Dirac quantisation condition
eg = 211"nli (2.16)
The presence of li in this condition makes obvious that we are discussing a

quantum effect. An immediate corrolary is that if a single monopole with
charge go exists then electric charge is quantised in units of 211"li/go.
In general when several electric and magnetic charges are present the
quantisation condition reads
(2.17)
where Nij E Z.
Exercise: Consider a dyon with electric and magnetic charge (e1' gl)
moving in the field of another dyon with charges (e2' g2). Redo the argu-
ment with the angular momentum to show that the electromagnetic angular
momentum is
(2.18)
which again implies that the appropriate quantisation condition here is
(2.19)
Another point of view is provided by the Dirac string singularity. As we

mentioned above the gauge potential is essential for the quantum theory.
jj = V x A implies for a smooth A that V·jj = O. However, for a point-like
magnetic monopole, V . jj '" 8(3) (T) so that the vector potential must have
a string singularity. To put it differently, the existence of a vector potential
implies that the magnetic flux emanating from a magnetic monopole must
have arrived in some way at the origin. This can be done by assuming that
we have an infinitely thin solenoid along say the z-axis which brings from
infinity the flux emanating from the monopole. This solenoid which smooths
out the string singularity can be shifted around by gauge transformations.
Thus, its position is not a physical observable and one should not be able
to measure it. This was the essence of the original argument of Dirac. The
phase acquired by a charge particle of charge e when transported around
the solenoid is given by
phase =e f Asolenoid . dl = eg = 211" integer (2.20)
wmch reproduces (2.16).

The upshot of all this is that we can consider including magnetic mono-
poles in electromagnetism. Then,
• The monopole charge satisfies the Dirac condition .
• The configuration is singular and has an unobservable string attached.
3. Non-abelian gauge theories

The ultraviolet behaviour of a U(l) gauge theory is singular (due to the
existence of the Landau pole which drives the theory to strong coupling).
It is believed that an IR U(l) gauge theory must be embedded in a sponta-
neously broken non-abelian gauge symmetry, in order to have regular UV
behaviour.
We will describe here the fate of Dirac monopoles in the context of
the spontaneously broken non-abelian theory. For the sake of concreteness
we will study the Georgi-Glashow model. It is an SU(2) Yang-Mills theory
coupled to scalars transforming in the adjoint. The Lagrangian is
L = ~F~vFa'JW + ~(DJI</»a(DJI</»a + V(</>) (3.1)

where
a 8vwJI-ef.
a - 8JI w v-
FJlV- a JI vc
abcWbw (3.2)
(DJI</»a = 8J1</>a - e f.abcW!</>c (3.3)
V(</» = ~(</>a4>a _ a2 )2 (3.4)
4
344 ELIAS KIRITSIS
The minimum of the potential is at 1¢1 2 = ¢a¢a = a2 • A vacuum is

described by a solution ¢g of the previous condition. A solution is charac-
terised by a non-zero three-vector ¢g with length a. This breaks the SU(2)
symmetry to U(1). The broken transformations rotate the vacuum vector
(Higgs expectation value). The unbroken gauge group corresponds to rota-
tions that do not change that vector. Obviously this group is composed of
the rotations around the vacuum vector and is thus a U(1).
The gauge boson associated to the unbroken U(1) symmetry (that we
will call the photon) is AJ.' = 4>g aW: . The electric charge (unbroken U(1)
generator) is given by
(3.5)
where T a are the 3 x 3 representation matrices of the adjoint of 0(3).

The particle spectrum of this spontaneously broken gauge theory is as
follows
I Particle I mass I spin I electric charge I

Higgs I~al 0 0
'Y I 0 I 1 0
w± I ea I 1 ±1
Exercise: Verify the above.
This theory has classical solutions (discovered by 't Hooft [31] and
Polyakov [32]) which are stable and carry magnetic charge under the unbro-
ken U(1) . One has to look for localised solutions to the equation of motion.
Far away the fields must asymptote to those of the vacuum. In particular
the Higgs field I¢I -+ a. We shift the potential so that at the minimum its
value is zero. We can write the Hamiltonian density as
The vacuum is characterised then by V(¢) = 0 as well as DJ.'¢a = 0, F:v = 0

Such a solution maps the two-sphere at infinity to the Higgs vacuum

manifold, which is given by three-dimensional vectors of fixed length. This
is also a two-sphere. The set of smooth maps from 82 --+ 82 are classified
topologically by their winding number, or their homotopy class and we have
7r2(82) = z.
The winding number is
(3.7)
The magnetic change of the soliton is related to the winding number thus:
47r
g=--w (3.8)
e
This seems not to be the minimal one required by the Dirac quantisation
condition. One would expect the minimal monopole charge to be 27r / e. This
is explained as follows: we can add fermions in the theory that transform in
the spin-1/2 representation (doublet) of SU(2). This would not affect the
monopole solution. On the other hand, now the fermions have U(l) charges
that are ±e/2 and they should also satisfy the Dirac condition. This can
work only if the minimal magnetic charge is 47r / e and this is the case.
The solutions with non-trivial winding at infinity must be classically
stable since in order to "unwrap" to a winding zero configuration they
must go through a singularity. Then their kinetic energy becomes infinite,
dynamically forbidding their decay.
To find the simplest w = 1 solution we use the most general spherically
symmetric ansatz
xa
¢a = - 2 H (aer) ,
e r
= 0 Wo (3.9)
.. x j
W a
t
= _E atJ _
er2
[1 - K(aer)] (3.10)
For large r, H --+ aer while K --+ o. At large distances the configuration
for AJ.! (the unbroken U(l) gauge field) is exactly the same as for a Dirac
monopole. One would ask: what happened to the Dirac string? This can
be seen as follows: with a singular gauge transformation we can map the
Higgs field that winds non-trivially at infinity, to one that does not. Due
to the singular gauge transformation the gauge field now acquires a string
singularity [31].
Exercise: Show that in the limit of large Higgs expectation value a --+
00 we recover the Dirac Monopole.
346 ELIAS KIRITSIS
We can also construct dyon solutions (as was first done by Julia and
Zee [33]) by allowing wg to be non-zero: wg = ~J(aer).
By manipulating the energy density of a soliton we can derive the fol-
lowing bound for its mass:
(3.11)
where e is the electric charge while 9 is the magnetic charge. This bound is
known as the Bogomolny'i bound and it is saturated when the potential is
vanishing.
In particular, the mass of the monopole in that case is given by M = a 9
and saturates the Bogomolnyi bound. Remembering the Dirac quantisation
condition, 9 = 47r/e we obtain M = 47ra/e. The mass of the W± bosons
also saturates the Bogomolnyi bound: M = a e. In perturbation theory,
e « 1, the W-bosons are much lighter than the monopoles.
Particles and solitons saturating the Bogomolnyi bound are called Bogo-
molnyi-Prasad-Sommerfield states or BPS states for short. We have seen
that the W-bosons and monopoles are BPS states in the case of zero po-
tential.
The simple model discussed above can be generalised to Yang-Mills
theories with any simple group G coupled to Higgs scalars that break the
group to a subgroup H. The vacuum again is specified by V(<p) = 0, DJ.L<p =
O. Taking the commutator [DJ.L<p, DII<PJ = FJ.LII<P we find that the unbroken
subgroup H is specified from Fp,II<Po = o. If the model does not have extra
global symmetries or accidental degeneracies then the vacuum manifold is
isomorphic to G /H. There are non-trivial monopole solutions if 7r2 (G / H)
is non-trivial. From the exact sequence
(3 .12)
one can compute the relevant homotopy group. We have 7r2(G) = 0 for
all G. When G does not contain U (1) factors 7r1 (G) = 0 as well so that
7r2(G/H) = 7r1(H) . Thus, there is a winding number (monopole charge) for
every unbroken U (1) factor.
Exercise: Show that the Standard Model does not have smooth mono-
poles.
In the general (G,H) case there is a generalisation of the Dirac quanti-

sation condition. This has been investigated by Goddard, Nuyts and Olive
[34] who found that the magnetic charges 9i take values in the weight lat-
tice A(H) of the unbroken group H. On the other hand the electric charges
qi take values in the dual of the weight lattice A*(H). Then the Dirac
condition can be written as
e q. 9 = 271"N (3.13)
with NEZ. The dual of the weight lattice is the weight lattice of the dual
group H* : A*(H) = A(H*). H determines the electric charges while H*
determines the magnetic charges. Moreover, (H*)* = H.
For H=SO(3) we have the Dirac quantisation condition e 9 = 471". The
dual group is H* = SU(2) with quantisation condition e9 = 271". For SU(N),
the dual group is SU(N)/ZN .
At this point we can describe the M ontonen - Olive conjecture [35].
A gauge theory is characterised by two groups Hand H*. There are two
equivalent descriptions of the gauge theory. One where the gauge group is
H, the conserved (Noether) currents are H-currents, while the H*-currents
are topological currents. In the other the gauge fields belong to the H*
group, the Noether currents are now the topological currents of the previous
description and vice versa. Moreover the coupling qln in the original theory
is replaced by 91n in the magnetic theory. Since 9 rv lie, this conjecture
relates a weakly coupled theory to a strongly coupled theory. It is not
easy to test this conjecture. Some arguments were given for this conjecture
originally. For example the monopole-monopole force was calculated and
was dual to the charge-charge force . , the conjecture cannot be true in
a general gauge theory. In the example of the Georgi-Glashow model the
massive charged states W±-bosons have spin 1 and duality maps them to
monopoles with spin O. One can bypass this difficulty by adding fermions
to the model. Fermions can have zero modes and thus give non-trivial spin
to monopoles making the validity of the conjecture possible. We need to
make monopoles with spin 1. On the way, there will be monopoles also with
spin 0 and 1/2. This way of thinking leads to N=4 supersymmetric Yang-
Mills theory as the prime suspect for the realization of the Montonen-Olive
conjecture.
4. Duality, monopoles and the O-angle

We have seen that for dyons the Dirac quantisation conditions reads
(4.1)
Let us consider a pure electric charge (q, 9) = (qO, 0) and a generic dyon
(qm,9m). Applying (4.1) we obtain qo9m = 271"n so that the smallest mag-
netic charge is 9min = ~~. Consider now two dyons with the minimum
348 ELIAS KIRITSIS
magnetic charge (ql,9min) and (q2,9min). Applying (4.1) again we obtain,
ql - q2 = nqo (4.2)
This is a quantisation condition, not for the electric charges but for charge
differences.
If we assume that the theory is invariant under CP
(q ,9) -+ ( -"
q 9) E -+ E , B -+ - B (4.3)
then the condition (4.2) has two possible solutions: q = n qo or q =

(n + ~) qo.
Gauge theories have a parameter that breaks CP: the O-angle. The ad-
dition to the Lagrangian is
(4.4)
Where NEZ in the integer valued, topological Pontryagin (or instanton)

number. Physics is periodic in the O-angle: 0 -+ 0+271" since eiS ' = eiS e27fiB =
eiS .
Exercise: Show that the theory is CP-invariant only for 0 = 0,71".
In the presence of the O-angle there is an "anomalous" contribution to

the electric charge of a monopole [36J
(4.5)
For a general dyon, one obtains from the Dirac condition
Oe 471")
(q,g)= ( ne+ 271"m, 7 m (4.6)
where n, m E Z. It can be seen that (4.6) verifies (4.2). We can obtain a

useful complex representation by defining
Q = q + ig = e (n +m [2: + i~;]) = e(n + mr) (4.7)

where we have defined the complex coupling constant
(4.8)
In this notation the Bogomolnyi bound becomes
M 2: ae In + mrl (4.9)
5. Supersymmetry and BPS states

We start with a brief review of the representation theory of N-extended
supersymmetry in four dimensions. A more complete treatment can be
found in [37].
Supersymmetry is a symmetry that relates fermions to bosons and vice
versa. Its conserved charges are fermionic (spinors). For each conserved
Weyl spinor charge we have one supersymmetry. In general we can have
more than one supersymmetry (extended supersymmetry).
The most general anticommutation relations the supercharges can sat-
isfy are [38]
I J IJ -I -J _ . -IJ I -J _ IJ '"
{QQ,Qfj}=€QfjZ ,{Qu,Qjj}-€ufjZ ,{QQ,Qu}-tS 2aQU P""
(5.1)
where ZIJ is the antisymmetric central charge matrix. It commutes with
all other generators of the super-Poincare algebra.
The algebra is invariant under the U (N) R-symmetry that rotates Q, Q.
We begin with a description of the representations of the algebra. We will
first assume that the central charges are zero .
• Massive representations. We can go to the rest frame P '" (M, 0). The
relations become
Define the 2N fermionic harmonic creation and annihilation operators
AI = _l_QI AtI = _l_Q-~ (5.3)

Q J2M Q' Q J2M Q'
Building the representation is now easy. We start with the Clifford vacuum
In), which is annihilated by the A~ and we generate the representation
by acting with the creation operators. There are (2::) states at the n-th
oscillator level. The total number of states is L~~o e::),
half of them being
bosonic and half of them fermionic. The spin comes from symmetrization
350 ELIAS KIRITSIS
over the spinorial indices. The maximal spin is the spin of the ground-states
plus N.
Example. Suppose N=1 and the ground-state transforms into the [j]
representation of SO(3). Here we have two creation operators. Then, the
content of the massive representation is [j] ® ([1/2] + 2[0]) = [j ± 1/2] + 2[j].
The two spin-zero states correspond to the ground-state itself and to the
state with two oscillators.
• Massless representations. In this case we can go to the frame P I"V
(-E, 0, 0, E). The anticommutation relations now become
{Q;, Q~} = 2(2: ~) 6IJ , (5.4)
the rest being zero. Since Q~, Q~ totally anticommute, they are represented
by zero in a unitary theory. We have N non-trivial creation and annihilation
operators AI = Q{j2v'E, At I = Q{j 2v'E, and the representation is 2N_
dimensional. It is much shorter than the massive one. Here we will describe
some examples (with spin up to one) that will be useful later on. For N=l
supersymmetry we have the chiral multiplet containing a complex scalar
and a Weyl fermion, as well as the vector multiplet containing a vector
and a majorana fermion (gaugino). In N=2 supersymmetry we have the
vector multiplet containing a vector, a complex scalar and two Majorana
fermions, as well as the hyper - multiplet, containing two complex scalars
and two majorana fermions. Finally in N=4 supersymmetry we have the
vector multiplet containing a vector, 4 majorana fermions and six real
scalars .
• Non-zero central charges. In this case the representations are massive.
The central charge matrix can be brought by a U(N) transformation to
block diagonal form 1,
0 Zl 0 0
-Zl 0 0 0
0 0 0 Z2
0 0 -Z2 0 (5.5)
0 ZN/2
-ZN/2 0
and we have labelled the real positive eigenvalues by Zm, m = 1,2, ... ,N/2.
We will split the index I --+ (a, m): a = 1,2 labels positions inside the 2 x 2
blocks while m labels the blocks. Then
{Q~m, Qr} = 2M6CXCt6abamn , {Q~m, Qr} = Znecx/3eab6mn. (5.6)
lWe will consider from now on even N.

Define the following fermionic oscillators
A a:m =.J2
1 [Qlm
a: + €a:/3 Q2m]
/3 '
Bm
a:
1 [Qlm
=.J2 a: - €a:/3
Q2m]
/3'
(5 7)
.
and similarly for the conjugate operators. The anticommutators become
{A~,A3} = {A~,B3} = {B;:,Bg} = 0, (5.8)
{A~, Abn } = 6a:{36 mn (2M +Zn) , {B;:, B1n} = 6a:{36 mn (2M -Zn). (5.9)
Unitarity requires that the right-hand sides in (5.9) be non-negative. This
in turn implies the bound
(5.10)
which turns out to be no other than the Bogomolnyi bound. Supersymme-

try in this sense "explains" the Bogomolnyi bound: it is essential for the
unitarity of the underlying theory.
Consider 0 ~ r ~ N /2 of the Zn's to be equal to 2M. Then 2r of the
B-oscillators vanish identically and we are left with 2N - 2r creation and
annihilation operators. The representation has 22N -2r states. The maximal
case r = N /2 gives rise to the short BPS multiplet whose number of states
are the same as in the massless multiplet. The other multiplets with 0 <
r < N /2 are known as intermediate BPS multiplets.
BPS states are important probes of non-perturbative physics in theories
with extended (N 2': 2) supersymmetry. The BPS states are special for the
following reasons:
• Due to their relation with central charges, and although they are
massive, they form multiplets under extended SUSY which are shorter than
the generic massive multiplet. Their mass is given in terms of their charges
and Higgs (moduli) expectation values.
• They are the only states that can become massless when we vary
coupling constants and Higgs expectation values.
• When they are at rest they exert no force on each other.
• Their mass-formula is supposed to be exact if one uses renormalised
values for the charges and moduli. 2 The argument is that quantum correc-
tions would spoil the relation of mass and charges, and if we assume un-
broken SUSY at the quantum level there would be incompatibilities with
the dimension of their representations.
• At generic points in moduli space (space of couplings and Higgs ex-
pectation values) they are stable. The reason is the dependence of their
2In theories with N ~ 4 supersymmetry there are no renormaIisations.

352 ELIAS KIRITSIS
mass on conserved charges. Charge and energy conservation prohibits their

decay. Consider as an example, the BPS mass formula
(5.11)
where m, n are integer-valued conserved charges, and 1" is a complex mod-

ulus. We have derived this BPS formula in the context of the SU(2) gauge
theory. Consider a BPS state with charges (mo, no), at rest, decaying into
N states with charges (mi' ni) and masses Mi, i = 1,2,···, N. Charge con-
servation implies that mo = L~l mi, no = L~l ni. The four-momenta of
the produced particles are (JMl + pt,Pi) with L~lPi = O. Conservation
of energy implies
N N
Mmo,no = L:JMl +pt ~ L:Mi (5.12)
i=l i=l
Also in a given charge sector (m,n) the BPS bound implies that any mass
M ~ Mm,n, with Mm,n given in (5.11). From (5.12) we obtain
N
Mmo,no ~ L:Mm;,n; , (5.13)
i=l
and the equality will hold if all particles are BPS and are produced at
rest (Pi = 0). Consider now the two-dimensional vectors Vi = mi + 1"ni
on the complex 1"-plane, with length Ilvil12 = Imi + ni1"12. They satisfy
Vo = L~l Vi· Repeated application of the triangle inequality implies
Ilvoll ~ L IIVil1 (5.14)

i=l
This is incompatible with energy conservation (5.13) unless all vectors Vi

are parallel. This will happen only if 1" is real which means when e =
00 a highly degenerate case. For energy conservation it should also be a
rational number. Consequently, for 1"2 finite, the BPS states of this theory
are absolutely stable. This is always true in theories with more than N> 2
supersymmetry in four dimensions. In cases corresponding to theories with
8 supercharges, there are regions in the moduli space, where BPS states,
stable at weak coupling, can decay at strong coupling. However, there is
always a large region around weak coupling where they are stable.
6. Duality in N =4 super Yang-Mills theory

The four-dimensional quantum field theory with maximal supersymmetry
is the N=4 Yang-Mills theory.3 The action ofN=4 Yang-Mills is completely
specified by the choice of the gauge group G (that we will assume simple
here). As pointed out in a previous section, the only N=4 multiplet with
spin at most one is the vector multiplet. The particle content is a vec-
tor multiplet in the adjoint of the gauge group containing a vector, four
fermions and six scalars. There is an SU(4) '" 0(6) global symmetry (the
R-symmetry). The supercharges transform in the 4, as well as the fermions,
while the scalars transform in the 6 (vector of 0(6)). The kinetic terms of
various particles as well as their couplings to the gauge field are standard.
The Lagrangian is
LN=4 = - 4~2 Tr [F~vF~v + XiJfxi + D~<PaD~<Pa + Yukawa terms+

(6.1)
t t] + 327rf)
+[<Pa, <Pb][<Pa, <Pb] 2 Tr
-
FF
The minima of the scalar potential are given by [<Pa, <p~] = 0 and they are
solved by a scalar belonging in the Cartan(G).
Exercise: Show that for a generic Higgs expectation value in the Cartan
of G, the gauge group G is broken to the abelian CartanG.
This is the generic Coulomb phase where the massless gauge bosons
are Nc photons, where Nc is the rank of G. The massive W-bosons are
electrically charged under the Cartan( G). Their masses saturate the BPS
bound and they are 1/2-BPS states (the shortest representations, as short
as the massless). There are also 1/2-BPS 't Hooft -Polyakov monopoles in
the theory.
The N=4 1/2-BPS mass formula is
(6.2)
3More than four supersymmetries in four dimensions imply the existence of spins
bigger than one and thus non-renormalisability. Such theories are good as effective field
theories.
354 ELIAS KIRITSIS
with T = 2()7r + i~.

9
¢ is the vev of the Higgs, while n, m are the integers
specifying the electric and magnetic charges respectively.
We will further set G = SU(2) for simplicity. The generalisation to
other groups is straightforward.
N=4 super yang-Mills for any gauge group is a scale invariant the-
ory. Its ,B-function is zero non-perturbatively. Moreover its low-energy two-
derivative effective action has no quantum corrections (even beyond per-
turbation theory). This does not imply, however, that the theory is trivial.
Correlation functions are non-trivial and it is an open p'"oblem to com-
pute them exactly (apart from some three point functions protected by
non-renormalization theorems).
Here the monopoles are in BPS multiplets similar to those of the W-
bosons and the Montonen-Olive duality has a chance of being correct. For
8 = 0 it involves inversion of the coupling constant 9 -+ "; as well as
interchanging of electric and magnetic charges n -+ m, m -+ -no If this is
combined with the periodicity in 8: 8 -+ 8+21l" we obtain an infinite discrete
group, SL(2,Z). It can be represented by 2 x 2 matrices with integer entries
and unit determinant
( aC db) , ad - bc = 1 , a, b, c, d E Z (6.3)
The associated transformations act as
(6.4)
There are two generating transformations: T -+ T + 1 (periodicity in 8) and

strong-weak coupling duality T -+ -l/r.
Exercise: Show that the BPS mass formula is invariant under the
SL(2,Z) duality.
Can we test Montonen-Olive duality? There are some further indications

that it is valid:
• In perturbation theory we have states with electric charge ±1 (the
W-boson multiplets). Then SL(2,Z) duality predicts the existence of dyons
with charges
(~ ~) (~) = (~) (6.5)

where the greatest common divisor of a,c is one, (a, c) = 1. All such dyons
must exist, if M-O duality is correct. For example, we have seen that the
(0,1) state, the magnetic monopole, exists in the non-perturbative spec-
trum. On the other hand no (0,2) monopole should exist, but the dyon
(1,2) should exist. This is a subtle exercise in geometry and quantum me-
chanics: one has to show that an appropriate supersymmetric quantum me-
chanical system on a non-trivial quaternionic manifold (the moduli space of
dyons with a given magnetic charge) has a certain number of normalisable
ground states. This in turn transforms into the question of existence of cer-
tain forms in the moduli space. This test has been performed successfully
for magnetic charge two [39] and the general case in [40] .
• There is a relatively simple object to compute in a supersymmetric
quantum field theory, namely the Witten index. This amounts to doing the
path integral on the torus with periodic boundary conditions for bosons
and fermions. On such a flat manifold the result is a pure number that
counts the supersymmetric ground states. If however, the path integral is
performed on a non trivial compact or non-compact manifold with super-
symmetry preserving boundary conditions, then the Witten index depends
non-trivially both on the manifold and the coupling constant T. The Witten
index for N=4 Yang-Mills was computed [41] on K3 and on ALE manifolds
and gave a result that was covariant under SL(2,Z) duality.
• In string theory, the M-O duality of N=4 super-Yang Mills is equiva-
lent to T-duality (a perturbative duality of string theory that is well under-
stood) via a string-string duality that has had its own consistency checks.
At this point we should consider the question whether it makes sense
to expect that we can have a way to prove something like M-O duality. In
order for this question to be meaningful, there must be an alternative way
of defining the non-perturbative (strongly coupled theory). Duality can be
viewed as a different (independent) definition of the strong coupling limit
and in that case it makes sense to ask whether the two non-perturbative
definitions agree. Unfortunately for supersymmetric theories we do not have
a non-perturbative definition. The obvious and only such definition (lattice)
breaks supersymmetry and remains to be seen if it can be used in that vein.
Montonen-Olive duality can be viewed as a (motivated and possibly in-
complete) definition of the non-perturbative theory. As with any definition
it must satisfy some consistency checks. For example if a quantity satisfies
a non-renormalization theorem and can be thus computed in perturbation
theory, it should transform appropriately under duality, etc. In all cases of
duality in supersymmetric field and string theories we are checking their
consistency rather than proving them.
356 ELIAS KIRITSIS
7. N=2 supersymmetric gauge theory

The two relevant N=2 massless multiplets are the vector multiplet and the
hypermultiplet. Here we will consider the simplest case: pure gauge theory,
with vector multiplets only. Hypermultiplets can also be accommodated but
we will not discuss them further here. The vector multiplet (A~, [X0, 1/J 0], AO)
contains a vector, two majorana spinors and a complex scalar AO all in the
adjoint of the gauge group.
The renormalisable N=2 Lagrangian is
LN=2 = ;2Tr [-~F~/lF~/I + (D~A)tD~A - ~[A,At]2 - i1/Ju~Dif;

(7.1)
-ixjj~D~x - iV2[1/J,X]At - iV2[7,li, x] A] + 3:7r2TrF~/lF~/I
This defines the ultraviolet theory. The theory is asymptotically free and
it :Hows to strong coupling in the infrared. The minima of the potential
are as before: A must take values in the Cartan of the gauge group. The
values at the Cartan are arbitrary (:Hat potential) and are moduli of the
problem. Put otherwise, there is a continuum of vacua specified by the ex-
pectation values of the Higgs in the Cartan. A non-zero (generic) Higgs
expectation value breaks the gauge group to the Cartan, U(I)Nc and we
are in the Coulomb phase. The GjU(I)Nc vector multiplets become mas-
sive (W-multiplets) and are BPS multiplets of N=2 supersymmetry since
they have the same number of states as the massless multiplets. There are
monopoles as usual since 7r2(GjU(I)Nc) = ZNc.
From now on we specialise to G=SU(2) to avoid unnecessary complica-
tions. Other groups can be treated as well.
The fundamental question we would like to pose here concerns strong
coupling. We have mentioned that the theory is asymptotically free. If one
is interested in physics at low energy then he has to solve a strong coupling
problem. As we will see, supersymmetry here will help us to solve this
problem. The end result will be the exact two-derivative Wilsonian effective
action at low energy. Obviously, the low-energy effective action is something
easy to calculate in an IR-free theory since one can use perturbation theory
(e.g. QED).
The Wilsonian effective action at a scale Eo is constructed by integrating
out degrees of freedom with energy E 2 Eo.
Going a bit back we can ask: what is the low-energy effective action for
the N=4 super Yang-Mills discussed in the previous section, in the Coulomb
phase. We have seen that the W-bosons are massive with masses"" l¢oI2.
If we are interested in energies smaller than their mass we can integrate
them out. The low-energy theory will contain only the photon multiplets
with possible extra interactions induced by the massive particles in the

loops. It turns out, however, that N=4 supersymmetry protects the two-
derivative effective action from corrections due to quantum effects (even
beyond perturbation theory). The most important part in the m, the two-
derivative action, again describes free photon multiplets with no additional
interactions. Moreover it is known that the four-derivative terms (like F4
terms) obtain corrections only from one loop in perturbation theory (in
four dimensions).
We would like to solve the same problem in the N=2 gauge theory, where
the two-derivative effective action does get quantum corrections from mas-
sive states. In this theory, the W-multiplets are massive with BPS masses
m 2= IAI2 where A is the third component of the non-abelian scalar which
parametrises the moduli space (a copy of the complex plane). We would
like to integrate out the W-bosons and find the effective physics for the
photon multiplet for energies well below the W mass IAI. The effective ac-
tion will of course be of the non-renormalisable type, a fact acceptable for
an effective theory. The low-energy effective action will contain a photon,
two photinos and a complex scalar A.
There are two special points in the space of vacua (moduli space) .
• A = O. Here the gauge symmetry is enhanced to SU(2), since the
W-bosons become massless .
• A -+ 00. This is the abelian limit and as we will see we can trust
perturbation theory in that neighbourhood of moduli space.
An important point to make is that we do not expect the N=2 super-
symmetry to break. Consequently, the effective field theory could be one
of the most general N =2 theories with a single vector multiplet. The most
general such (non-renormalisable) action is known. It depends on a single
unknown function F known as the prepotential which is a holomorphic
function of the complex scalar A. We summarise it below.
ff2F[ -41 Fl'v F I' v + DI'ADI' At ] +Re 8

Leff '" Im 8A2
2F 1 -
8A2 32 nF I'v F I'v +fermions
(7.2)
As obvious from above Im ~~ is the inverse effective coupling while Re ~~
is the effective O-angle.1t is obvious that if we manage to find F(A) we have
completely determined the low-energy effective action.
Classically (at the tree level) F(A) = ~T A2 reproduces the classical
(UV) coupling constant T. The prepotential F(A) will have both perturba-
tive and non-perturbative corrections (coming here from instantons).
An important ingredient of the effective U(l) theory is the value of
the central charge (that determines the BPS formula) as a function of the
358 ELIAS KIRITSIS
modulus A:
(7.3)
where n e, nm are integers that determine the electric and magnetic charges
respectively. Here we see an example where the central charge receives quan-
tum corrections (since F does) but the mass equality M = IZI for BPS
states still remains valid. This happens because the mass is also renor-
malised as to keep the BPS relation valid.
At tree level we have
*.
(7.4)
We will define the dual Higgs expectation value AD == Then we have

the following M-O-like SL(2,Z) duality:A +-+ AD, ne +-+ n m.
We need a better coordinate than A on the moduli space. The reason is
that A is not gauge invariant. The Weyl element of the original SU(2) gauge
group acts as A -+ -A. Thus, a gauge-invariant coordinate is u = A2/2.
At A = u = 0 we have gauge symmetry enhancement U(l) -+ SU(2).
7.1. THE FATE OF GLOBAL SYMMETRIES
An N=2 supersymmetric theory has a U(2) = U(l) x SU(2) (global) R-

symmetry that rotates the two supercharges. The various fields of the vector
multiplet transform as follows:
I Particle I U(l) I SU(2) I

AI-' 0 I singlet I
x,'I/J 1 I doublet I
A 2 I singlet I
The U(l) R-symmetry has a chiral anomaly, which means that it is

broken by instanton effects. For a gauge group SU(N)4 an instanton has a
zero mode for each left fermion in the fundamental and 2N zero modes for
a fermion in the adjoint. Here our fermions are in the adjoint. In order to
obtain a non-zero amplitude in an instanton background we need to soak
the fermionic zero modes, and that can be done by inserting the appropriate
'We will do this analysis for general N although eventually we will be interested in
N=2.
number of fermion operators in the path integral. We thus obtain that the
simplest non-vanishing correlator is
2N 2N
G= (11 X(i) II 1f;(i)) =F 0 (7.5)
i =l i =l
G has U(l) charge 4N and transforms under a U(l) transformation eia as

G ~ ei4Na G. This implies that since G =F 0, the U(l) symmetry is broken
to Z4N . The unbroken global symmetry is SU(2) x Z4N. However, the
centre of SU (2) (that acts as (1f;, X) 4- - ( 1f;, X)) is contained in Z4N · We
conclude that the global symmetry is (SU(2) x Z4N)/Z2. When we have a
non-zero Higgs expectation value A, the global symmetry breaks further.
For example in the SU(2) case u rv A2 has charge 4 under Z8 so that Z8
breaks to Z4. For SU(2) this is the end of the story and the unbroken global
symmetry is (SU(2) x Z4)/Z2. The broken Z8 acts as u ~ -u.
Exercise: Find the unbroken global symmetry for G=SU(3), SU( 4).
7.2. THE COMPUTATIONAL STRATEGY
We need to calculate the holomorphic prepotential :F( u) in order to de-

termine the exact effective action. The central idea is that if we know the
singularities and monodromies of a holomorphic function then there is a
concrete procedure that reconstructs it.
The strategy is [42] to find the singularities and monodromies of :F(u) .
• Use perturbation theory to study the singularity at u ~ 00.
• Use physical arguments and local SL(2,Z) duality to determine the
behaviour at the other singular points.
• Use math techniques to reconstruct :F(u).
Classically the only two singular points are A 4- 00 and A 4- 0 where
we have gauge symmetry enhancement and the U(l) effective theory breaks
down.
7.3. PERTURBATION THEORY
An important ingredient in perturbation theory is that the two-derivative

effective action obtains corrections only at one loop (in the presence of un-
broken N=2 supersymmetry) . The argument is simple. The (anomalous)
360 ELIAS KIRITSIS
eff
g
m E
Figure 1. The running coupling past a threshold.
divergence of the R-current 0ISJ'it PF belongs to the same N=2 super-

I'V
multiplet with the trace of the energy-momentum tensor T lSv . Classically

the theory is scale invariant and TlSv is traceless. , quantum effects break
scale invariance and in the quantum theory the trace is proportional to the
,8-function of the theory. On the other hand the axial anomaly obeys an
Adler-Bardeen non-renormalization theorem that specifies that in a given
scheme (the Adler-Bardeen scheme) it receives quantum corrections at one
loop only. Unbroken N=2 supersymmetry implies that this is also true for
the ,8-function of the theory and consequently for the prepotential. We are
left with a one-loop calculation to do.
The one-Ioop,8- function in field theory is given by the following formula
o
J.L OJ.L geff (J.L) == ,8(g) (7.6)
1 1
g2(J.L) = 95 - 1
871'2 ~ bi log
(J.L 2 + m?)
A2 I
(7.7)
where the ,8-function coefficients are given by
bi = (_1)28Q2 (112 - 82) (7.8)
Here 8 is the helicity and Q is an appropriately normalised generator of the

gauge group. A boson contributes 1/12, a Weyl fermion 1/6 while a vector
contributes -11/12. The summation is over all particles, with masses mi.
Expression (7.7) is approximate at the thresholds (when J.L comes near to
one of the masses mi) but very accurate elsewhere.
Assume for simplicity that there is only one particle with mass m con-
tributing to the ,8-function. The following behaviour of the effective cou-
pling can be seen from (7.7):
• For J.L > > m there is logarithmic running .

• For J.L « m the coupling "freezes" at the value g-2(m) = g02 -
~ log ~. This is reasonable since for energies lower than m all co.ntri-
butions of the particle have been integrated out . Consequently there IS no
further running of the coupling. This behaviour is portrayed in Fig. 1.
The massive particles we are integrating out are two massive vector
multiplets. Their mass is m = IAI. The contribution of a single vector
multiplet to the ,6-function coefficient is bv = 2112 + 4~ - 2l~' = -1. The
electric charge is 1 so that in total b = -2 Q2 = -2. Since we integrate out
all energies above the mass of the particles the effective coupling for energy
below IAI is frozen to
1 1 2 IAI2
-2-
geff
= 2"
go
+ -8
7r
2 log A2 (7.9)
We can absorb go into A (dimensional transmutation) and rewrite
1 1 IAI2
-2- = 4 2 log A2 (7.10)
geff 7r
This must come from a holomorphic prepotential F(A) so that
ImF"(A) 1 IAI2
--4-7r--'---~ = 2 log -A-2-
-47r- (7.11)
The solution is
. A2
F(A) = _2 A 2 log /1.. 2 (7.12)
27r
By allowing A to be complex, we can absorb into it the classical 8-angle.

To one loop,
8efflone-loop = 4(Arg(A) - Arg(A)) (7.13)
In what region of the moduli space can we trust perturbation theory?
This can be seen from Fig. 1. Now m = IAI. By taking IAllarger and larger
while keeping A (the UV coupling) fixed, the effective coupling freezes at
lower and lower values. Thus, in the neighbourhood of A = 00 perturbation
theory is reliable.
As can be seen from the one-loop prepotential there are two singu-
larities: A = 0 and A = 00. The singularity at A = 00 we trust since
perturbation theory is a good guide there. This is not true for the one at
A = 0 where the theory is strongly coupled. Can this be the only singu-
larities of the prepotential? The answer is no, for the following reasons: A
holomorphic function with two singularities on the complex plane, and a
362 ELIAS KIRITSIS
logarithmic cut at 00 (remember that we trust this) is unique and given by

the one-loop result.
On the other hand, this is incompatible for two reasons .
• For smaller values of A, the coupling constant becomes negative .
• The one-instanton contribution to the ,a-function had been computed
before and found to be non-zero.
The only way out is to assume that F(A) has more singularities on the
complex plane.
7.4. SINGULARITIES AND MONODROMY
Consider the complex function f (z) = ..;z. If we encircle once the origin,
z -+ e21ri z, then f(e 21ri z) = - f(z). The function does not return back
to itself. This is a signal that the point z = 0 is a singular point for the
function, in this case the start of a branch cut. The behaviour of a complex
function or a set of functions after transport around a point (singularity)
is called the monodromy. In general a set of functions, transported once
around the singular point Zo return to a linear combination of themselves.
We write
(7.14)
The matrix M depends on the singular point, and is called the monodromy
matrix at that point. Monodromy has a topological character. The mon-
odromy matrices do not change under smooth deformations of the contour.
Non-smooth deformations include the contour crossing another singular
point.
This matrix is important because it plays an essential role in the lliemann-
Hilbert problem: if we know the position of the singularities and the mon-
odromy around each one, of a set of holomorphic functions, then we can
reconstruct them uniquely.
If we want to be a bit more careful then we will realize that F( a) is not
really a function. We have seen earlier that SL(2,Z) duality interchanges the
derivative of F, AD with A. The relevant holomorphic objects to consider
are the pair A and AD viewed both as functions of the good coordinate
u = A 2 /2 . If we make a circle around u = 0, then u -+ e21ri u and A -+ -A.
2iA ( A
AD = F (A) = ---;- log A + 2
I 1) (7.15)
When A -+ - A then
AD -+ F I (-A) = -2iA 1)
- ( log --A + - = -AD + 2A (7.16)
1r A 2
Q
8
Figure 2. The global monodromy condition
The monodromy around u = 0 is given by
(7.17)
Similarly, the monodromy around u = 00 is
(7.18)
The two matrices satisfy Mo Moo = 1. This is a general property of mon-

odromy. If we have a number of singularities on the sphere then the associ-
ated monodromy matrices satisfy IIi Mi = 1. The proof of this is sketched
in Fig. 2. We start with a number of independent contours that we can
deform until we obtain a single one that we can shrink to zero on the back
side of the sphere.
As we mentioned above, if we only have two singularities then the per-
turbative result is the whole story. We had argued though that instanton
corrections are non-trivial. We will analyse now their expected form. From
364 ELIAS KIRITSIS
the one-loop running we have
The k-th instanton contribution is proportional to exp[-k~] '" (1fk.

This breaks the U(1) R-symmetry as expected (A is charged). We can
restore the U(1) symmetry if we allow A to transform with charge 2. Then
the exact prepotential is expected to have the following form,
i A2 + A
Y(A) = 27r log A2
2 00
~ Ck
(A)4k
A (7.19)
One needs to calculate the coefficients Ck.

We have seen that we need more singularities than the ones we have
observed in perturbation theory. The possible meaning of such singularities
would be that they are due to states that become massless at that particular
point of the moduli space. This would signal the breakdown of the effective
theory, since we have integrated out something very light. There are two
possibilities; the particles that become massless are in vector multiplets
or in hypermultiplets. The guess of Seiberg and Witten is that only the
second case is correct. First we have an abundance of non-perturbative hy-
permultiplets, namely monopoles and dyons that could in principle become
massless at strong coupling. There are various arguments that indicate that
it is implausible that vectors become massless [42].
One extra constraint is that singularities that appear on the sphere
except the points A = 0 and A = 00 must appear in pairs. The reason
is that if a singularity appear at u = Uo then by the broken R-symmetry
it must be that also u = -Uo is a singularity. The minimal number of
singularities we need is three. Since A = 00 is a singularity, we must also
have a pair of singularities in the interior of the moduli space. In that case,
the classical singularity at A = 0 must be absent non-perturbatively. These
assumptions can be verified a posteriori.
We put two extra singularities, one at u = A2 and another at u = -A 2
(this can be thought of as a non-perturbative definition of A. A natural
guess for the particle that becomes massless at u = A2 is that it is the
monopole. , there are monodromy constraints that must be satisfied and
we must take them into account.
We will assume that some dyon becomes massless at a given point of the
moduli space and try to compute the monodromy matrix. The low-energy
theory around the singularity must include the very light dyon. Then we
would like to compute the local coupling by computing a one-loop diagram
where the dyon is going around the loop. This is not obvious how to do. It
is duality at that point that comes to the rescue.
7.5. THE DUALITY MAP
We will need the following identities in four dimensions
(7.20)
The quadratic action can be written as
S = _1_Im
3~
I r(a)(F + iF)2 = _1_Im
~~
I r(a)(2F2 + 2iF F) (7.21 )
If we want to consider F as an independent variable we must explicitly

impose the Bianchi identity dF = O. This we can do by adding an extra
term in the action
(7.22)
Integrating over the vector V~ gives a 6-function that imposes the Bianchi
identity. !':l.S can be rearranged as follows
11 {)~VlI€~lIIX7FIX7=--
!':l.S=-- 1 /FFD=-6-Re
- 11- (FD-iFD)(F+iF)-
8~ 8~ 1 ~
(7.23)
where FD = dV.
Exercise: The action S + t!:..S is quadratic in F. Integrate out F to

obtain the dual action:
- 1 I( - r(a)1) (FD + zFDFD)

S = 16~ Im
2 . - (7.24)
The above indicates that near the point where the monopole becomes
massless the low-energy theory contains the photon as well as the monopole.
By doing a duality transformation as above we can write the low-energy
theory in terms of the dual photon. With respect to it the monopole is
electrically charged, and if the coupling is weak one can use normal pertur-
bation theory.
We can choose a local coordinate A(p) = C(u-uo) around the point Uo
where the monopole becomes massless. The mass of the monopole behaves
366 ELIAS KIRITSIS
as M2 '" IA(p)12. The theory around that point is IR free (since it is photons
plus charges). As we go go close to the singularity, M -+ 0, perturbation
theory (in the dual variables) becomes better and better. The ,a-function
coefficient due to a charged hypermultiplet is bH = 412
+ 4! = 1. This
implies that locally the prepotential is
(7.25)
and the dual coordinate
AD(p) == -8F
- = -iA
- [ log-_
A2(p)
- +1 1 (7.26)
8A(p) 27r A2
Now we can go around Uo: u - Uo -+ (u - uo)e 27ri . Since A(p) = C(u - uo)
we obtain that A(p) -+ A(p). Also from (7.26) we obtain AD(p) -+ AD(p) +
2A(p). The monodromy matrix is
( AD)
A ,
-+ M(O,l) (AD)
A (1 2)
= 0 1 (AD)
A (7.27)
We are interested in the monodromy matrix in the original variables. We

have performed a 7 -+ -1/7 transformation in order to map the monopole
to an electric charge. We now have to invert this transformation. We find
(7.28)
Exercise: Consider a point where the (ne, n m ) dyon becomes massless.

By doing the appropriate duality transformation it can be treated as an
electrically charged particle, whose local monodromy we have already com-
puted. Invert the duality map to compute the monodromy matrix and show
that
(7.29)
If the dyon (n,m) becomes massless at u = A2 and (n ' , m/) at u = _A2

then we must have
(7.30)
This can be solved to find the following solutions
I (m,n) I (l,n) I (-I,n) I (-I,n) (l,n)

I (m',n') I (l,n-l) I (1,-n-l) I (-I,n+l) I (-I,-n+l) I
The simplest solution is obtained for m = m' = 1, n = 0, n' = -1. It
can be shown that it is the only consistent solution.
So we are almost finished. We know all singular points of the holomor-
phic frame (A(u),AD(U)) and the associated monodromy matrices. It re-
mains to use them to solve for A(u), AD(U). The answer is that A(u), AD(U)
are given by the two periods of an auxiliary torus. The effective coupling
constant T is given by the modulus of the torus. The periods of this torus
vary as we change the modulus u.
The explicit solution can be written in terms of hypergeometric func-
tions [42]
vf2
A(u) = --;-
/1
-1
";x -
dx ~ =
V2(1 + u) F (-2'12'11; 1 +2)
U
u (7.31)
vf2 r ";x-u i (11 l-U)

AD(U) = --;- 11 dx x2 -1 = 2(u -1) F 2' 2,2; -2- (7.32)
F( a, b, C; x) is the standard hypergeometric function. We have set A = 1. It

can be put back in on dimensional grounds. Once we have (7.31,7.32) we
can compute the effective coupling T as
A'
T(U) = A~ (7.33)
where the prime stands for the u-derivative.

The positions of the three singularities coincide with the positions where
the auxiliary torus degenerates (a cycle shrinks to zero).
In conclusion we have managed to calculate the exact low-energy two-
derivative effective action of an SU(2) N=2 gauge theory. This theory has
one parameter: the ultraviolet value of the coupling constant or equivalently
A. For IAI > > A the effective theory is weakly coupled and perturbation
theory is reliable. However, here we have controlled the effective theory for
IAI ~ A where the effective coupling is strong.
The appearance of the torus in the Seiberg-Witten solution can be ex-
plained naturally by embedding the gauge theory into string theory [3].
8. Monopole condensation and confinement

Consider a U(I) gauge theory (QED) which is spontaneously broken by the
non-zero vacuum expectation value of a (electrically charged) scalar field
368 ELIAS KIRITSIS
(Higgs). This is precisely what happens in normal superconductors. The

appropriate Higgs field is a bound state of electrons (Cooper pair) with
charge twice that of the electron. Electric charge condenses in the vacuum
(= the Higgs gets an expectation value) and the photon becomes massive.
A well known phenomenon in such a phase is the Meissner effect. Mag-
netic fields are expelled from the superconducting bulk. There is only a thin
surface penetration which goes to zero with the distance from the surface
as e- m r. This is because the photon is massive in the superconductor and
the parameter m is no other than the photon mass. Thus, magnetic flux is
screened inside a superconductor.
Consider now introducing a magnetic monopole inside the supercon-
ducting phase. The magnetic flux emanating from the monopole will be
strongly screened and will form a thin flux tube. IT there is an anti-monopole
around, the flux tube will stretch between the two. At low energies such
a flux tube is elastic and behaves like a string: the energy is propor-
tional to the stretching. There is a linear potential between a monopole-
antimonopole pair inside a superconductor. This means that magnetic mono-
poles are permanently confined in the superconducting (electric Higgs)
phase. As we try to pull them apart we must give more and more en-
ergy. Eventually when we have given energy greater than that required for
a monopole-antimonopole pair to materialise from the vacuum the string
will break and we will end up with two bound states instead of separated
magnetic charges.
The dual phenomenon was argued to be the explanation for the perma-
nent confinement of quarks [43, 44]. Here, we need a magnetically charged
object (monopole) to get an expectation value in the vacuum (magnetic
condensation). The ensuing dual Meissner effect will confine the electric
flux and the electric charges. Although this mechanism remains to be seen
if it is responsible for confinement in QCD, we will argue here following
[42] that it does explain confinement in an N=1 gauge theory that we will
obtain by perturbing the N=2 gauge theory we have considered so far.
We would like to softly break the original N=2 SU(2) gauge theory
to N= 1. For this we split the N=2 vector multiplet into an N= 1 vector
multiplet (A~, Xa) and in a N=1 chiral multiplet ~ == (1/Ia, Aa). We will add
a superpotential V rv m Tr~2 to make ~ massive. At energies much smaller
than m, ~ decouples and the theory is N=l SU(2) super Yang-Mills which
is an asymptotically free theory. We would expect confinement, a mass gap
and breaking of chiral symmetry (which here is Z4 as discussed before).
Consider the superpotential V = m Tr~2 /2 = m U where U is the N=1
superfield whose scalar component is our coordinate u. IT one goes through
the same procedure of integrating out massive states one would get an extra
potential in the low-energy effective theory. It can be shown [42] that the
induced superpotential is identical with the ultraviolet one. Consider now

the effective theory near the point where the magnetic monopole becomes
massless. To smooth out the effective field theory we must include the
monopole multiplet in our effective action. The superpotential has an N=2
piece that gives the mass to the monopole'" IADI as well as the N=l
superpotential
(8.1)
where M,.tV! denote the two N=l components of the monopole hypermulti-
plet. To find the ground state of the effective field theory we must minimise
the potential: dW = 0
In - du -
v 2M M + m dAD = 0 , AD M = AD M = 0 (8.2)
At a generic point AD -I- 0 the solution to the second equation (8.2)

is < M >=< .tV! >= 0 . Substituting in the first equation we obtain
du/dA D = O. This can never be true since u is a good global coordinate on
the moduli space.
The only stable vacuum in the neighbourhood exists for AD = O. From
(8.2) we find that the monopoles have a non-trivial expectation value
< M >=< M- >= ~

--u'(O)
v'2
(8.3)
It can be checked from the exact solution that u'(O) is negative.

What we have found is: a magnetically charged scalar has acquired a
vacuum expectation value. It breaks the (magnetic) U(l) gauge group and
generates confinement for the electric charges. The fate of the massless
fields is as follows: the U(l) vector multiplet acquires a mass from the
Higgs mechanism while the monopole hypermultiplet is "eaten up" by the
vector multiplet. The upshot is that everything is now massive and the
mass gap is proportional to the Higgs expectation value in (8.3). This value
is non-perturbative.
A similar analysis around the point where the dyon becomes massless
gives similar results. There we have a realization of the oblique confinement
of 't Rooft. The theory we started with has two ground states, and this is
explained by the chiral symmetry being broken from Z4 --+ Z2.
9. Epilogue of field theory duality

We have seen that in an N =4 supersymmetric field theory we expect an
exact duality symmetry that interchanges weak with strong coupling.
370 ELIAS KIRITSIS
In the context of N = 2gauge theories the solutions of Seiberg and Witten

do generalise to arbitrary gauge groups[45, 46] as well as the inclusion of
"matter" (hypermultiplets). The exact effective description can always be
found both in the Coulomb as well as in the Higgs phase. There can be also
mixed phases but they can be treated similarly.
The situation becomes more interesting in the context of N=1 gauge
theories. A general non-renormalisable N=1 field theory is specified by three
functions of the chiral fields:
.: The Kahler potential K(zi, zi) this is a real function and determines
the kinetic terms of the chiral fields. Their geometry is that of a Kahler
manifold with metric Gil = 8;E%K .
• The superpotential W(zi). It is a holomorphic function ofthe chiral
fields and has R-charge equal to two. The potential can be written in terms
of the superpotential, and the Kahler metric (we ignore D-terms) as
(9.1)
• The gauge coupling function J(z). It is also a holomorphic (and gauge
invariant) function. Its imaginary part determines the gauge coupling con-
stant while its real part the O-angle.
In the N=1 case, unlike the N=2, we do not have full control over
the two-derivative effective action. We can determine , the holomorphic
superpotential. Assuming smoothness of the unknown Kahler potential,
knowledge of the exact superpotential specifies uniquely the minima and
thus the ground-states of the effective field theory. Here again the strategy
is to start from a remormalisable, asymptotically free gauge theory and
find the superpotential in the low-energy (strongly-coupled) effective field
theory as well as the ground states.
The N=1 SU(Nc ) gauge theory was studied [47] coupled to NF chiral
multiplets in the fundamental and its complex conjugate. We will briefly
present some of the most interesting results. For more details the interested
reader should consults more extensive reviews on the subject [1] as well as
the original papers [47].
When N F > Nc + 1 the theory has a dual "magnetic" description: the
dual gauge group is SU(NF - N c ) and the charged matter is composed of
N F flavours of quarks as well as a set of N~ gauge singlet "mesons" Mij '
These meson superfields are supposed to correspond to the mesons of the
original theory
1 ..
Mij = - q'if' (9.2)
J.L
where J.L is a dynamical scale.
Moreover there is an electric-magnetic type duality between the two
theories (Seiberg duality) which can be expressed as a relationship between
i
\ i
\
\
------~
\
AF IRF
)
//
Q,
_
: "'--........ n --------------/('
:
~
..__.__.-...........
I "-....... "
.., . / I
I ............ __ ._ . .......... _._-
A A
A increases ->strong coupling A decreases ->strong coupling
A decreases -> weak coupling A increases -> weak coupling
Figure 3. Running coupling for asymptotically free and infrared free theories
their A parameters as follows:
(9.3)
where Nc = NF - N c·
The one-loop ,a-function co~:fficient of the original theory is b = NF-3Nc
while that of the dual theory b = 3Nc - 2NF.
In the range Nc + 1 ~ N F < ~Nc the electric theory is asymptotically
free while the magnetic theory is IR free. The magnetic theory can be
used to describe the low-energy dynamics in a weak coupling regime. The
relation (9.3) can be seen to indicate that when the electric coupling is
strong the magnetic coupling is weak and vice versa (see Fig. 3). In the
region ~ Nc :::; N F ~ 3Nc both theories are AF and they flow to a non-
trivial fixed point in the IR.
An interesting and important question is: what can be done when there
is no supersymmetry or when supersymmetry is broken? Duality ideas seem
that they can handle the softly broken case [48]. However, calculations in
the broken theory can be trusted once the supersymmetry breaking scale
is much smaller that the dynamical scale(s) of the theory.
10. Introduction to String Theory

String theory was born in 1968 [49] as a candidate theory to describe the
dual properties of hadrons. It has been superseded by QeD, and reemerged
in 1976 [50] as a candidate theory of gravity and all other fundamental
372 ELIAS KIRITSIS
--.~ 0
Figure 4. String theory versus field theory diagrams
interactions. In 1984 it acquired a big impetus [51] due to the tightness of

constraints [52] on possible consistent theories.
String theory postulates that the fundamental entities are strings rather
than point-like objects. , from a large distance a string can be viewed as a
point-like object. At distances well above the string length Is string theory
is well approximated by field theory. String perturbation theory resembles
field theory perturbation theory, (diagrams fatten, see Fig. 4) but has also
different properties in the UV.
• Closed string theory predicts gravity. If one quantises free strings in a
flat background, a spin-two massless state can be found in the spectrum. It
has gravitational type interactions and can be identified with the graviton.
• It is a theory that is UV-finite. In some sense string theory can be
though of as a collection of an infinite number of quantum fields with a
"smart" UV cutoff of the order of the string scale Ms.
• String theory provides a consistent and finite theory of perturbative
quantum gravity.
• Existence of spacetime fermions in the theory implies supersymmetry.
• String theory unifies gravity with gauge and Yukawa interactions nat-
urally.
• The theory has no free parameters (apart from a scale Is = Ms-l) but
many ground-states (vacua). The string coupling constant is related to the
expectation value of a scalar field, the dilaton 98 = e<. The continuous
-'- ---------,..,
-
RI 0-
-r R-O
Figure 5. A circular extra dimension can be invisible when R is small.
parameters of various ground-states are always related to expectation values

of scalar fields. The string tension is 1-;2 .
• String theory was defined in perturbation theory until '95. Since then
duality ideas allowed us to explore string theory beyond perturbation the-
ory, indicate that the theory is unique, suggest the existence of a most
symmetric eleven-dimensional theory and provided new tools and ingredi-
ents for its study. We do not as of now have a complete non-perturbative
formulation of the theory.
• Superstrings live in ten or less large (non-compact) dimensions. A
topical question is: How come we see four large dimensions today? Kaluza
and Klein long time ago, suggested how the two could be compatible. The
idea is that some dimensions can be small and compact and can thus avoid
detection (see Fig. 5). We will present here a five-dimensional example for
the sake of simplicity. We consider a massless scalar in five dimensions,
with mass-shell condition p2 = o. Consider now the fifth coordinate to be
a circle with radius R. Then the components of the momentum along the
fifth directions .is ~uantised. This is obtained from the periodicity of the
wavefunction eap5 :C under shifts x 5 -t x 5 + 2piR. One obtains P5 = i
where m E Z. The five-dimensional mass-shell condition can be written as
2
2 ::J2 m
Po - P = R2 (10.1)
Equation (10.1) indicates that from the four-dimensional point of view, this
five-dimensional massless scalar corresponds to an infinite tower of particles
(called Kaluza-Klein states) with masses M = lJil.
When our available
energy E << 1/ R no experiment can produce a KK particle. Moreover,
their loop effects are suppressed. Thus, at E < < 1/R the extra dimension
is unobservable. For E ~ 1/R the effects of the KK particles and thus
the extra dimension become visible. For example, if such a particle feels
the standard model forces then this implies an upper bound on the radius
which is of the order R '" 1O-2o m '" (lOTeV)-l [53]. On the other hand
it is quite surprising that if only gravity lives in five dimensions, then the
374 ELIAS KIRITSIS
radius can be as large as R ,...., 1O-4 m [54, 55] without contradicting current
experimental data [56].
The ten-dimensional part of the action governing low-energy gravity
(below the string scale) is
1
810 ,...., - 2[8
98 8
! d 10 x v. r-;;:;
-G R + ... (10.2)
If six dimensions are compactified on a manifold with volume V6[~ then the
four-dimensional Einstein action obtained from (10.2) will be
84 ,...., - V6
9 s2[2s
! d4 x vr-::
-9 R + ... (10.3)
from where we can read the four-dimensional Planck mass
(lOA)
The following regimes are important:

• For energies below the string tension, E < M s , strings cannot have
their vibrational modes excited. Their dynamics is associated with their
centre of mass motion and can be thus described by standard field theory.
On the contrary, for E > Ms the stringy modes can be excited and the
physics departs sensibly from the field theory behaviour .
• For energies E < < Mp gravity is very weak and can be neglected at
the microscopic level. Its quantum effects are unimportant. On the other
hand, for E 2: Mp gravity becomes strong, th quantum gravitational effects
cannot be treated perturbatively and it is not known how to handle the
theory in this case.
There are three possibilities concerning the hierarchy of scales:
- Mp ,...., Ms. This is the conventional scenario, where gauge fields come
from the perturbative closed string sector. Both stringy as well as
quantum gravitational phenomena are far removed from near future
experiments and experimental signals of string theory are obscured by
the huge disparity in scales. For this to happen, 9s ,...., O( JV6) . There
are two possibilities in perturbation theory: 9s ~ 0(1) and a compact
manifold of Planck size, or an hierarchically small coupling constant
and a sub-Planckian compact manifold. In the second case, this can be
mapped via T-duality to a string theory with a large volume compact
manifold.
- Ms < < Mp. In this case stringy phenomena can be visible at low
energy, hopefully at near future accelerator experiments, while quan-
tum gravity remains out of current reach. For this to work out, there
are two possibilities. First, a hierarchically small coupling constant

and a Planckian size manifold. Second, gs :::; 0(1) and a large com-
pact manifold. In this case the threshold of KK excitations is of order
MKK '"[76 «Ms. The first signal will be production of KK states
f"V
V6
before stringy effects are visible.
- Mp < < Ms. This necessitates (in perturbation theory) a sub-Planckian
compact space which will be mapped via T -duality to a different string
ground state.
11. T-duality
Classical strings behave very different from point particles at distances of
order the string length, ls. A characteristic feature is that closed strings can
stretch and wrap around a non-contractible cycle of a compact manifold.
Consider again the five-dimensional example with one direction being a
circle of radius R. The energy cost for a string wrapping n times around
the circle is given by
Ewrapping = (total length) x (string tension) = 21rn ~ (11.1)

s
Now, the mass-shell condition (10.1) is modified to

2 2
2 :=e m 2 2n
Po - P = R2 + 41T' R [4 (11.2)
s
A symmetry (a special case ofT-duality) is obvious in the mass formula

(11.2):
_ [2
R -+ R = 2:R ' m ++ n (11.3)
The physical content of this stringy symmetry is that we cannot distinguish
a circle with size smaller than the string length. The effective radius we
measure is always
(11.4)
When R is large the low lying excitations are the KK states. When R is
small, the low lying excitations are winding modes, that can be interpreted
as KK modes with a dual radius R. T-duality is a symmetry of string theory
valid order by order in perturbation theory.
The fact that the string cannot distinguish length scales that are smaller
that its size is no surprise. What is a surprise is that a circle with length
much smaller than the string length is equivalent to a macroscopic one.
376 ELIAS KIRITSIS
Classical strings at distances larger than the string scale, feel the stan-
dard Riemannian geometry. At smaller scales, the Riemannian concept
breaks down. The generalisation is provided by Conformal Field Theory
which could be viewed as an infinite-dimensional generalisation of Rieman-
nian geometry [57, 58] . This can have deep implications on the geometric
interpretation of strong curvature as well as early cosmological phenomena
[58].
12. A collection of superstring theories
Until recently we were blessed with an embarrassment of riches: we knew

five distinct, stable, consistent, supersymmetric string theories in ten di-
mensions.
Closed Strings
• Type-II strings. These are the most normal of all strings. They are
closed strings, with isomorphic left-moving and right-moving modes. There
are also fermionic oscillations responsible for the appearance of spacetime
fermions . They are Lorentz invariant in ten-dimensional flat space. There is
a subtle difference of "gluing" together the fermionic left and right movers.
This results in two distinct string theories:
- type IIA: This is a non-chiral ten-dimensional theory with N=2 space-
time supersymmetry. The low-energy effective field theory is type IIA
supergravity. Its bosonic spectrum contains the graviton, a two-index
antisymmetric tensor and a scalar (the dilaton) as well as a set of forms
(Ramond-Ramond states): a vector and a three-form.
- type lIB: This is a chiral, anomaly-free ten-dimensional theory with
N=2 supersymmetry. The low-energy effective field theory is type-lIB
supergravity. The bosonic spectrum contains the graviton, two-form
and dilaton (like the type IIA) but the Ramond-Ramond (RR) forms
are different: here we have a zero-form (scalar) , another two-form and
a self-dual four-form. One can make a complex number T = a + ie-4>,
by putting together the RR scalar (axion, a) and the dilaton (string
coupling constant, 98 = e4» Then, the effective type-lIB supergravity
is invariant under a continuous SL(2,R) symmetry which acts projec-
tively on T:
(
a
c d
b) E SL{2, R) , T -+
aT
CT
+b
+d (12.1)
The two two-forms transform as a doublet, while the Einstein metric

and the four-form are invariant. This is reminiscent of a similar situa-
tion in N = 4 super Yang-Mills theory. It is expected that the presence
of objects charged under the two-forms will break the continuous sym-
metry to a discrete subgroup, namely SL{2,Z).
+ ......
Figure 6. The first few diagrams for the propagator of a closed string theory
CJ
a) b) c) d)
Figure 7. The first few diagrams (with boundaries and unorientable surfaces) for the
vacuum energy of an open string theory: (a) Disk (b) Annulus (c) Moebius strip (d) Klein
bottle
Both type II strings cannot fit the fields of the Standard Model in perturba-
tion theory. This is partly due to the fact that gauge fields descending from
the RR sector have no charged states in perturbation theory and cannot
thus serve as Standard Model gauge fields .
• Heterotic strings. This is a peculiar type of string [59]. The idea is
that since left and right-movers are independent one can glue superstring
modes on the right (living in ten dimensions) and bosonic string modes
on the left (living in twenty six dimensions). The extra sixteen left-moving
coordinates are required by consistency to be compactified on the two pos-
sible even self-dual sixteen-dimensional lattices: the root lattice of Es x Es
or that of Spin(32)j Z 2 5. The low-energy effective field theory is N=1 D=10
supergravity coupled to D=1O super Yang-Mills with gauge group Es x Es
or SO(32). The bosonic spectrum is composed of the metric two-form and
dilaton, as well as the gauge bosons in the adjoint of the gauge group.
For all the closed string theories the structure of perturbation theory
is elegant: each order of perturbation theory corresponds to a computation
using the appropriate Conformal Field Theory on the associated Riemann
surface. The perturbative expansion is organised by the number of loops
(genus or number of handles of the associated Riemann surface), and there
is a single diagram per order. This includes (in the low-energy limit) the
contributions of N! distinct diagrams of field theory (see Fig. 6).
Open and Closed Strings: Type-I string theory. The theory contains
both closed and open unoriented strings. From the closed string sector we
5This is the root lattice of SO(32) augmented by one of the two spinor weights.
378 ELIAS KIRITSIS
0=11
ITfl» ",1
Heterotic eterotic
0=10 Type-! Type IIA
0(32) E"xE"
0(32)
" , T' "

T ', " , ' T
o
"" If T
o
"" ;/
o
Figure 8. Perturbative and non-perturbative connections between string theories
obtain N=l supergravity in ten dimensions while from the open string sec-
tor we obtain 80(32) super Yang-Mills. The structure of the perturbation
theory is more involved now since it involves both open and closed surfaces,
as well as both orient able and non-orientable surfaces. The first few extra
terms in the genus expansion are shown in Fig. 7.
13. Duality connections
We have seen that we have five distinct supersymmetric theories in ten

dimensions. Are they truly distinct or they form part of an underlying
theory?
In string perturbation theory there are two connections that are shown
in Fig. 8 with broken arrows.
Upon compactification to nine (or less) dimensions on a circle of radius
R the heterotic Es xEs and 0(32) theories are continuously connected. In
nine dimensions, we can turn-on Higgs expectation values6 and break the
gauge group. We have two limits in which we can go back to ten dimensions:
The first is to take R -+ 00. If we started with the 0(32) string we will
end up with the 0(32) string in ten dimensions. The other is R -+ o. You
remember that using T -duality R -+ 0 is still equivalent to a very large
circle. If we adjust appropriately the Wilson lines in this limit we end up
with the Es x Es string. This indicated that the two ten-dimensional theories
6These are scalars that come from the tenth components of the gauge fields in ten
dimensions. These expectation values are called Wilson lines.
Figure 9. Perturbative and non-perturbative connections between string theories
are not disconnected but corners in the same moduli space of vacua of a
single (the heterotic) theory (see figure 9).
A similar situation exists for the type IIA and type lIB theories. Al-
though they look very different (for example one is chiral the other is not)
once they are compactified to nine dimensions they are related by T-duality.
At R = 00 one recovers the ten-dimensional type lIA theory while at R = 0
we recover the ten-dimensional type lIB theory (figure 8).
If we go beyond perturbation theory we will find more connections [60,
61]. The key is to ask what is the strong coupling limit of the various ten-
dimensional string theories. The tools to investigate this question we have
already discussed in the field theory context: they are supersymmetry and
BPS states .
• The type-IIA theory contains point-like solitons (known today as DO-
branes) that are electrically charged under the RR gauge field (remember
no perturbative state has electric or magnetic charge under RR forms).
Their mass is given by
n
MDO = - , nEZ (13.1)
9s
where n is the electric charge. Since these are 1/2-BPS states we can trust
their mass formula also at strong coupling. We learn that at strong coupling
380 ELIAS KIRITSIS
---f----+ - - ---
d=10 d=10
R
Figure 10. The non-perturbative Es xEs heterotic string as a compactification of
M-theory on a interval
they become arbitrarily light. This tower of states reminds us of the tower
of KK states for large radius. This is not accidental: it was long known
that the action of ten-dimensional type-IIA supergravity could be obtained
by dimensional reduction of eleven-dimensional supergravity on a circle of
radius R. The KK states of the graviton have a spectrum like the one
in (13.1) and they are charged under the off-diagonal components of the
eleven-dimensional metric that becomes the RR gauge field. The precise
relation is
(13.2)
Thus, we expect that the strong coupling limit of type IIA theory is an
eleven-dimensional theory (named M-theory) whose low-energy limit is
eleven-dimensional supergravity [61]. Compactifying M-theory on a circle
we obtain type-IIA string theory.
• On the other hand compactifying M-theory on the orbifold Sl/Z 2
we obtain the Es xEs heterotic string theory. The string coupling and the
radius of the orbifold are still related as in (13.2). The orbifold is defined
by moding the circle out by the inversion of the coordinate (1 -t -(1. This
projects out the low-energy spectrum of M-theory to N=l ten-dimensional
supergravity. We also have two fixed points of the action of the orbifold
transformations: (1 = 0, 7r. These are fixed ten-dimensional planes, and as it
happens in perturbative string theory, there are extra excitations localised
on the orbifold planes. Anomaly cancellation indicated that each plane
should carry a ten-dimensional Es Yang Mills supermultiplet (figure to).

In the perturbative heterotic string (small R) the two planes are on top of
each other whereas they move apart non-perturbatively.
• The strong coupling limit of type-lIB theory is isomorphic to its weak
coupling limit. This is due to the fact that an SL(2,Z) subgroup of the
continuous SL(2,R) symmetry is unbroken and that includes the transfor-
mation that inverts the coupling constant .
• Finally, the two 0(32) theories, namely the heterotic and the type-l
are dual to each other. This means that the strong coupling limit of the
heterotic theory is the weakly coupled type-l theory and vice versa.
All these connections are summarised in figure 8 and the overall picture
is portrayed in figure 9. We learn that the five string theories are corners
in a moduli space of a more fundamental theory.
14. Forms, branes and duality

We have seen that the various string theories have massless antisymmetric
tensors in their spectrum. We will describe here the natural charged objects
of such forms and how electric-magnetic duality extends to them.
We will use the language of differential forms and we will represent a
rank-p antisymmetric tensor AJllJ12 ... Jlp by the associated p-form
Ap == AJllJ12 ... Jlp dx Jl1 1\ ... 1\ dx Jlp • (14.1)

Such p-forms transform under generalised gauge transformations:
Ap ~ Ap + d Ap - b , (14.2)
where d is the exterior derivative (d 2 = 0) and Ap - 1 is a (p - I)-form
that serves as the parameter of gauge transformations. The familiar case
of (abelian) gauge fields corresponds to p=1. The gauge-invariant field
strength is
Fp+l = d Ap. (14.3)
satisfying the free Maxwell equations
d* Fp+1 = 0 (14.4)
The natural objects, charged under a (p+l)-form Ap +1, are p-branes. A
p-brane is an extended object with p spatial dimensions. The world-volume
of p-brane is (p+l)-dimensional. Point particles correspond to p=O, strings
to p=1. The natural coupling of Ap+1 and a p-brane is given by
exp [iQp Jrworld - volume A p+1] = exp Q [i p ! AJlo ... JlpdxJlO 1\ ... 1\ dx JlP ] ,
(14.5)
382 ELIAS KIRITSIS
which generalises the Wilson line coupling in the case of electromagnetism.

This is the a-model coupling of the usual string to the two-index antisym-
metric tensor. The charge Qp is the usual electric charge for p=O and the
string tension for p=1. Qp has mass dimension p + 1. For the p-branes we
will be considering, the (electric) charges will be related to their tensions
(mass per unit volume).
In analogy with electromagnetism, we can also introduce magnetic char-
ges. First, we must define the analog of the magnetic field: the magnetic
(dual) form. This is done by first dualizing the field strength and then
rewriting it as the exterior derivative of another form 7 :
dA D - P - 3 = FD - p- 2 =* Fp+2 =* dAp+l , (14.6)
where D is the the dimension of spacetime. Thus, the dual (magnetic) form
couples to (D-p-4)-branes that play the role of magnetic monopoles with
"magnetic charges" QD-P-4.
There is a generalisation of the Dirac quantisation condition to general
p-form charges discovered by Nepomechie and Teitelboim [62, 63]. The ar-
gument parallels that of Dirac. Consider an electric p-brane with charge Qp
and a magnetic {D-p-4)-brane with charge QD-p-4. Normalise the forms
so that the kinetic term is ~ I* Fp+2Fp+2. Integrating the field strength
Fp+2 on a (D-p-2)-sphere surrounding the p-brane we obtain the total flux
 = Qp. We can also write
 r r
= JSD-p-2 *Fp+2 = JSD-p-3 A D - p- 3 , (14.7)
where we have used (14.6) and we have integrated around the "Dirac
string". When the magnetic brane circles the Dirac string it picks up a
phase ei c}QD-p-4, as can be seen from (14.5). Unobservability of the string
implies the Dirac-Nepomechie-Teitelboim quantisation condition
(14.8)
The type IIA string theory contains a one- and a three-form in the
RR sector. They couple electrically to a particle (DO-brane) a membrane
(D2-brane) and magnetically to a D6-brane and and D4-brane. Moreover
there is a non-propagating nine-form that couples to a D8-brane. There
is always the fundamental string that couples electrically to the two-index
antisymmetric tensor. Its magnetic dual is the NS5-brane.
In the type lIB theory we have a zero- two- and self-dual four-form.
The electric and magnetic branes are the D1, D3, D5 and D7 branes. The
is also a D-instanton (denoted also by D{-1)).
7This is guaranteed by (14.4).
Figure 11 . A D-brane with an attached string
These branes can be described as solitonic extended objects in the low-

energy supergravity theory. All are Ij2-BPS states and thus preserve half
of the original supersymmetry.
15. D- branes
There is an exact stringy description of the solitonic branes we have men-
tioned in the previous section (except the NS5-brane). They can be defined
as defects (walls) in spacetime where closed strings can end. A closed string
when moving always stays closed. When it interacts with a brane it can
open up and its end-points are forced to move on the brane (figure 11).
The fluctuations of such open strings are essentially the fluctuations of the
brane itself. They can be shown to carry the appropriate RR charge. Their
name derives from the Dirichlet boundary conditions obeyed by the open
strings attached to the brane.
The quantisation of the open strings on a Dp-brane gives a massless
spectrum that is that of maximal Yang-Mills supermultiplet in p+l dimen-
sions. It contains a single vector, 9-p scalars and the associated fermions.
Note that is is the dimensional reduction of an N=l Yang-Mills supermul-
tiplet in ten dimensions.
The p-brane has some obvious collective coordinates, namely its posi-
tion in the transverse (9-p )-dimensional space. The expectation value of the
9-p scalars are precisely these collective coordinates. They have no poten-
384 ELIAS KIRITSIS
tial since we can put a brane anywhere in the transverse space. There is an
effective action on the D-brane that describes its dynamics. It can be calcu-
lated from the string description. Since the D-brane is a 1/2-BPS state, its
world-volume action will be supersymmetric (N=1 in ten dimensions (D9),
or N=4 in four dimensions (D3) etc.) Moreover at low energies the action
must reduce to the super Yang-Mills action. The effective action is
(15 .1)
:= I
d?+1x e-¢ [1 + FabFab +8ax l aaxI + ...]
where the induced metric on the brane is
(15.2)
This action describes in general the dynamics of the brane modes as well
as their coupling to the bulk string fields. It is non-linear and comes under
the name of Dirac-Born-Infeld (DBI) action. Note that the energy per unit
volume is proportional to 1/9s. For normal solitons the dependence is 1/92 .
An interesting phenomenon happens when we have many coincident D-
branes. As can be seen in figure 12, if we label the branes by 1 and 2 then
there are four possible strings:l-l, 2-2, 1-2, 2-1. Each will give rise to a
massless Yang-Mills multiplet (if the branes coincide in transverse space).
It turns out that the gauge symmetry now is non-abelian, namely U(2).
This can be inferred from the fact that a string end-point on the brane acts
like an electric charge for the gauge field coming from the string with both
end-points on the brane. Note that the scalars Xl that we had interpreted
as the coordinates of the D-brane have now become 2 x 2 matrices. This is
an interesting realization of ideas concerning the quantisation of spacetime
(the coordinates becoming non-commuting operators).
What happens when by keeping the branes parallel we separate them a
distance I in the transverse space (figure 12)? The two strings (1-2, 2-1) are
now stretched by a distance I and give a shift in the energy CT I where CT is
the string tension. The two gauge bosons associated to them are no longer
massless: they have a mass CT 1. In the effective theory on the branes, this
is the ordinary Higgs effect. The U(2) Yang Mills has a potential
The minimum is when Xl are diagonal matrices (the Cartan of U(2))
I
X min = (xi x~0)
0 (15.3)
Figure 12. Two parallel D-branes and the various open string fluctuations
The vacuum expectation values x{, x~ have the interpretation of the coor-
dinates of the transverse position of the two branes. The two off-diagonal
gauge fields acquire a mass proportional to IXI - x21 in accordance with our
expectations.
The generalisation is straightforward for N parallel branes. The gauge
group is U(N). The overall U(l) corresponds to the centre of mass position
while the SU(N) describes the internal dynamics. In the generic vacuum
the branes are all separate and the gauge symmetry is broken to U (1 )N .
The state of affairs has some important messages
• The spacetime positions of branes correspond to the vacua of the
world-volume Yang-Mills theory.
• The fluctuations of the D-branes are the fluctuations of the Yang-Mills
theory.
• The interaction of the brane with the bulk supergravity fields is pro-
vided by the word-volume couplings. For example the interaction with the
spacetime metric GJ-IV is obtained from the following modification of the
induced metric in (15.2)
(15.4)
etc.
There are numerous applications of the previous observations:
• Geometric/brane Engineering. The strategy here is the following. We
put together branes so that we construct our favourite gauge theory inc1ud-
386 ELIAS KIRITSIS
ing some matter content. We compactify some directions and take 98 -+ 0

to decouple gravity while keeping the gauge coupling fixed. We can then
study properties of the associated Yang-Mills theories from the spacetime
picture of the branes. The results include a derivation of the M-O duality for
N=4 super Yang-Mills theory, derivation of Seiberg-Witten type solutions
for N=2 theories as well as the Seiberg Duality for N=1 gauge theories.
• Black-hole state counting.
• Gauge theory/gravity correspondence.
16. Black-holes and D-branes

In the early seventies, culminating with the works of Bekenstein and Hawk-
ing it was realized that black-holes obey laws similar to ordinary thermo-
dynamics.
• They have entropy given by one quarter the area of the horizon in
gravitational units
1 A
SBH= - - (16.1)
4GN
• They radiate thermal radiation with temperature
(16.2)
where K, is the specific gravity on the horizon.

• They satisfy all thermodynamic laws. The first
dM = TdS + work (16.3)
where work terms can be related to angular momentum, charge etc. The
second dS ~ 0 is also satisfied (by classical gravity) as well as the third.
The above observations create a clash with quantum mechanics known
as the "black hole information paradox" that can be summarised as follows:
Form a black hole from matter in an initially pure state. Let it evaporate
completely via thermal Hawking radiation. Then the whole system has
transformed into a mixed state and this is is not permitted by quantum
mechanics. There have been many attempts to resolve this paradox till
today, but it is fair to say that the paradox still stands.
One can cook up a similar paradox with a star. The star is formed by
matter in a pure state that is eventually squeezed by gravity, heating up,
and radiating thermal radiation. Here , there is no paradox. We do know
that the initial correlations are encoded in the outgoing radiation which is
not exactly (only approximately) thermal. In order to argue this, we have
as a tool the microscopic statistical mechanics of all particles that form
a star. Without knowing the microscopic degrees of freedom one cannot

resolve this paradox.
The important question is: what are the microscopic degrees of freedom
responsible for the Bekenstein entropy (16 .1) of a black hole?
Until a couple of years ago this question went unanswered. Here we will
show that string theory gives the microscopic degrees of freedom responsible
for the entropy.
Consider a particle of mass M. IT the Schwarschild radius of the parti-
cle is much bigger than the fundamental gravitational length ip then the
particle can be viewed as a black-hole. This will happen if the particle has
a mass much larger than the Planck mass.
String theory has many such states with masses M » Mp. Moreover
for large M their density grows exponentially as eC M. This implies that
their entropy is linear with the mass. However, the Bekenstein entropy for
black holes grows quadratically with the mass (in four dimensions).
There is already a problem with the comparison though. The masses of
string states obtain generically large quantum corrections. The mass enter-
ing the Bekenstein formula is the physical mass (after quantum corrections
have been taken into account) whereas the mass in the density is the bare
mass. The way out is to look for states that are protected from quantum
corrections. These are precisely the BPS states in supersymmetric theories.
We would like to put together many of those and create a smooth black
hole.
Consider a charged black hole, with charge Q. Then the BPS bound is
M ~ IQI. If the black hole is extremal (BPS) then M = IQI and it has
zero Hawking temperature. It is stable as expected from supersymmetry.
It must , have a horizon of finite area (or equivalently, finite macroscopic
entropy)
The simplest example of that sort can be constructed in five dimensions
in the type IIB theory. In order to have an extremal black-hole with non-
zero horizon area, it must carry three distinct charges, Q1, Q5 and Qo (in
four dimensions we need four).
The first is to find the supergravity solution, compute the area and then
the Bekenstein entropy. This can be done and the result is
(16.4)
The second step is to construct the black-hole out of a collection of

elementary states, in all possible ways, matching its charges. We need to
use D-branes (wrapped on cycles so as to give point-like objects in D=5). So
we consider type-lIB theory compactified on C4 X S1 (where C4 can be T4 or
K3), to five dimensions. Consider a bound state formed out of Q5 D5-branes
compactified around C4 x S1 and Q1 Dl strings wrapped around S1. IT we
388 ELIAS KIRITSIS
consider the volume of C4 to be much smaller than that of 8 1 , the world-

volume theory on the branes can be reduced to 1+1 dimensions (time+81 ) .
We can still add some fluctuations without breaking supersymmetry. We
can consider left-moving waves with "energy", Qo in the (1+1)-dimensional
theory. They do not break all of supersymmetry so we are still considering
an extremal configuration. It can be shown that the numbers Qo, Ql, Qs
correspond to gauge charges of the bound-state.
Now we would like to take the string coupling to be small 9s < < 1 so
that gravity is weakly coupled. We would like also to have a bound-state
that is macroscopic: its Schwarschild radius should be much larger than
the Planck scale. For this to happen, the parameters 98Qi > > 1 for all Qi .
This implies that although we have suppressed closed string interactions
the interactions of the D-brane modes which have coupling constants 9sQi
are strong.
We have two distinct limits:
• 98Qi < < 1. Here the bound state is point-like, but we can count states
since we are dealing with weakly coupled gauge theory.
• 9sQi > > 1. Here the gauge theory is strongly coupled and we cannot
compute microscopically. In this region, the bound-state is a macroscopic
black-hole.
Supersymmetry bridges the gap between the two regions. The states
we will be counting will be unpaired BPS states. Unbroken supersymmetry
guarantees that the counting at weak coupling holds true at strong coupling.
To count at weak coupling we note that the effective two-dimensional
theory is a hyper-Kahler <i-model with central charge c = 6(Q1Q + 1) ~
6Q1 Qs for large charges. We need the density of states at level Qo and this is
given by the Cardy formula in Conformal Field Theory: p '" exp[21TvQoc/6]
which gives for the entropy
8microscopic = logp = 21TVQOQ1QS + .. . = 8Bekenstein + ... (16.5)
where the ellipsis stands for subleading contributions. The two results agree.
The subleading contributions have been compared too and agree. In gravity
the correction comes from R4 terms in the effective action.
This is the first example known where a microscopic counting of black-
hole degrees of freedom agrees with the semiclassical, gravity result. Similar
agreement is found for more general extremal and near extremal black holes
in five and four dimensions.
A further question concerns a more involved calculation: that of the
Hawking radiation emission rate. This rate, known as a greybody factor
(since it encodes also the interaction of the outgoing radiation with the
gravitational field) has a non-trivial dynamical content and it is not pro-
tected by supersymmetry in general. However, we have good reasons to
Figure 13. The process of Hawking evaporation in the D-brane black-hole
believe that when supersymmetry is slightly broken (near extremality) the

calculation is reliable at strong coupling [64, 65].
In our five-dimensional example, near extremality means to add a small
admixture of right-moving waves in the two-dimensional conformal field
theory. When a right and a left-moving wave scatter, closed strings can be
produced: this is the Hawking radiation [66] (see figure 13).
Needless to say that the greybody factors calculated from the D-brane
approach agree with the gravitational calculation [67, 68].
There are two open problems. Extend the above to black holes far away
from the supersymmetric limit (Schwarschild for example). And solve the
information paradox. We should say that at weak coupling the description
is manifestly unitary. The question is: where the horizon is formed and
whether there is an associated phase transition. The problem is open.
17. Gauge theory I gravity correspondence
We have discussed earlier in these lectures an intuitive picture of colour

confinement: electric flux is forced into thin flux tubes, that behave like
strings with energy rising linearly with the distance and responsible for
permanent confinement of quarks (that are attached at the ends of such
flux tubes). It is known since the early days of the liNe expansion, [69]
that such a string description of the strong interactions becomes a good
approximation when the number of colours Ne -+ 00. In particular we
would expect that different classes of large-Ne gauge theories to give rise
to different effective string theories.
390 ELIAS KIRITSIS
The running coupling of SU(Ne) gauge theory is given by
a 11 gfM 5
iL aiL gy M = - 3 Ne 167r2 + O(gy M) (17.1)
In order to have a regular expansion for the coupling we define the 't Hooft
coupling
(17.2)
The standard large-Ne limit is defined as Ne --t 00 keeping A fixed

and small. To leading order, SU(Ne) is indistinguishable from U(Ne), the
propagator of the gauge field A~ can be given as two lines, ~ne for the
fundamental index i and one for the antifundamental index j and thus
the typical perturbative diagrams automatically fatten up to become two-
dimensional surfaces. Moreover, the weighting factor for the diagrams scales
a N;-2 g where 9 is the genus of the tW<rdimensional surface spanned by the
fattened graph. This is the first and important indication that the theory
in this limit is described by some string theory although the nature of this
string theory, despite many attempts over the years remained elusive.
The general expansion structure of (perturbative) observables is
L L
00 00
O(A,Ne) = N;-2 g Cg(A) , Cg(A) = Cg,M AM , A« 1

g=O M=O
(17.3)
The would be string coupling constant is given by gs = liNe.
This string cannot live in four flat dimensions. For one reason it is well
known that no string theory in flat space has Lorentz invariance apart from
ten or twenty six dimensions. As Polyakov [70] argued one needs at least an
extra dimension in order to match symmetries (like the zig-zag symmetry)
with the Wilson loop of the gauge theory.
The new idea in this direction is: A D-brane carries a gauge theory on its
world-volume. This theory in a certain regime can reproduce gravitational
effects (and vice versa) . By now this is not a complete surprise, since already
in our discussion of black hole entropy in the previous section, we have seen
examples of this. In fact, this correspondence was inspired by the black-
hole investigations. The crucial point here is two dual descriptions of a
single object. This object is a bound-state of many branes. At weak t'Hooft
coupling, its physics is best described by weakly coupled gauge theory. At
strong coupling the bound-state is macroscopic and self-gravitating and can
be well described by gravity.
We will look more carefully in the simplest example of this correspon-
dence. Consider a solitonic (gravitational) solution of type-lIB supergravity,
the black D3-brane:
ds
2= - f(r)dt2 + di
yfH(T)
2
+ sqrtH(r) (dr 2
f(r) + r 2d0 52) (17.4)
The three coordinates i and time describe the word-volume of the D3-brane
while r and the five angles on S5 parametrise the transverse space.
L4 r4
H(r) = 1 + 4" ' f(r) = 1 - ~ (17.5)
r r
There is also a non-zero spherically symmetric self-dual four-form C4 [71].
The position of the horizon is r = roo L4 = gsNli, where N is the
charge (number) of the D3 brane. Supergravity is a good approximation
in this background when the curvature is small compared to the string
scale L > > Is. This implies gsN > > 1. On the other hand g8 < < 1 so
that gravitational loops are suppressed. The limits are compatible when
N~oo.
We will keep the "distance" with units of energy U = r II; fixed and we
will take the Is ~ 0 limit in order to decouple unnecessary string modes.
This is the near-horizon limit of the black D3-brane.
() gsN
H r ~ 14 4 (17.6)
aU
and the metric becomes (we set ro = 0)
(17.7)
which is the metric of AdSs xSs. This metric has the symmetry 0(2,4) xO(6)
as well as maximal supersymmetry (32 supercharges). AdSs has a bound-
ary at infinity U = 00, that it is isomorphic to four-dimensional Minkowski
space and can be reached at finite time from any point of the interior.
We will consider now the same object as a collection of N parallel D3-
branes. We have two kinds of excitations, open strings (fluctuations of the
D3-branes and closed strings,( bulk fluctuations). The effective action will
have the form
S D3 = Sbulk + SBrane + Sinteraction (17.8)
We will take the limit Is ~ 0 keeping dimensionless parameters fixed. The
bulk theory in this limit becomes free gravity (since gravity is IR free). The
same is also true for the interaction action that describes the interactions
of brane and bulk fields. The only non-trivial interactions that are left over
are the interactions of Sbrane namely those of N=4 U(N) super Yang-Mills.
392 ELIAS KIRITSIS
In the previous supergravity description, as 18 -+ 0 the excitations at

r -+ 00 decouple from those near the horizon. This is due to a potential
barrier near the horizon that makes absorption cross sections to vanish with
vanishing energy as <Tabs'" w3 L8. For the excitations near the horizon, the
potential barrier keeps them from spreading out.
Moreover, a.s Is -+ 0 the fluctuations further away from the horizon are
described by free supergravity. Matching the two descriptions we obtain
the
Maldacena Conjecture [72]: N=4 D=4 SU(N) super Yang-Mills is
dual to type-liB string theory on Ad85 X 8 5 • The gauge theory descrip-
tion is weakly coupled when A < < 1 while the supergravity description
is insensitive to stringy data when A > > 1. The symmetries of the su-
pergravity theory match the conformal 0(2,4) symmetry of Yang-Mills a.s
well a.s its 0(6) R-symmetry. The extra enhanced supersymmetry is due to
conformal invariance. It should be stressed , that for this correspondence
supersymmetry is not important.
The precise form of the correspondence states [73, 74] that quantum cor-
relators in Yang-Mills match a.ssociated "S-matrix elements" in supergravity 8
Many strong coupling data of the gauge theory can be simply computed
in the supergravity picture. A typical example are the Wilson loops[75, 76],
but also the particles (glueball spectrum [77] etc. )
Supersymmetry can be broken in two ways: turn on temperature in a
higher dimension [78] or find non-supersymmetric brane solutions of the
effective equations [79].
Although this correspondence is a major step forward towards under-
standing gauge theory and gravity an important question still remains: find
the right QeD string!
18. Conclusions and outlook

We have gone a long journey through some major theoretical developments
of the past five years. In the context of field theories major progress has
been made towards understanding the strong coupling dynamics of super-
symmetric theories. The vacuum structure of N=l gauge theories or the
low-energy effective action of N =2 theories are some of the cornerstones of
the effort. A key ingredient in the above is electric-magnetic duality. Su-
persymmetric theories naturally admit the concept that seems to capture
some properties of the dynamics.
8Strictly speaking there are no S-matrix elements in AdS. , at tree level we can define
them using the usual procedure (equations of motion). However, they do not have the
traditional interpretation in terms of scattering, but as we saw, they have to do with
boundary correlators.
We have gone further and applied similar techniques based on dualities

to supersymmetric string theory. The outcomes are:
• We have learned that there is unique theory encompassing different
looking string theories.
• There is a most symmetric vacuum in eleven dimensions corresponding
to a theory coined M-theory whose low-energy limit is eleven-dimensional
supergravity. M-theory unifies many apparently dual descriptions in lower
dimensions.
• String theory contains many new objects, D-branes and other branes
that are essential for the consistency of the theory.
• D-branes provide a new and deep link between gauge theory and grav-
ity. They hint at quantisation of spacetime. They provide the microscopic
degrees of freedom responsible for black-hole entropy and might illuminate
the puzzles of quantum gravity.
• In certain low-energy limits they provide a link between gauge theory
and gravity that leads to supergravity (or stringy) descriptions of gauge
theories. The hope is that this will lead to a string theory for QCD.
There are many problems that are not yet solved. More relevant here
is supersymmetry breaking for duality treatments of strong coupling prob-
lems. Although there are some cases analysed , it is not known how much
of the non-perturbative treatments survive supersymmetry breaking. The
expectation is that for soft susy breaking and small susy breaking param-
eters, duality related non-perturbative techniques are still applicable. This
as well as strong breaking are open problems.
Another open problem is the search for the QCD string. Although the
results so far are negative in the context of the gauge theory/gravity corre-
spondence, there does not seem to be a reason forbidding its existence. It
maybe that the supergravity (or brane configuration) turns out to be too
complicated. In any case, it is one of the important problems of theoretical
high energy physics.
Finally, applications of the above to the physics of the Standard Model
and beyond will be a concrete way to emphasise the value of these devel-
opments.
Acknowledgements
The author wishes to thank the organisers of the 99' Cargese Summer
School on Particle Physics for hospitality and the students for creating a
stimulating atmosphere. This work was partially supported through a TMR
contract ERBFMRX-CT96-0090 of the European Union.
394 ELIAS KIRITSIS
References
1. K. Intriligator and N. Seiberg, Nuc!. Phys. Proc. Supp!. 45BC (1996) 1 hep-
th/9509066.
2. A. Bilal, hep-th/9601007.
3. W. Lerche, "Introduction to Seiberg-Witten Theory and its Stringy Origin", Nucl.
Phys. 55B [proc.Supp!.] (1997) 83; Fortsch. Phys. 45 (1997) 293; hep-th/9611190.
4. L. Alvarez-Gaume and S.F. Hassan, Fortsch. Phys. 45 (1997) 159 hep-th/9701069.
5. M.E. Peskin, hep-th/9702094.
6. R. Dijkgraaf, hep-th/9703136.
7. M. Shifman, Prog. Part. Nucl. Phys. 39 (1997) 1 hep-th/9704114.
8. L. Alvarez-Gaume and F. Zamora, hep-th/9709180.
9. P. Di Vecchia, "Duality in N=2,4 Supersymmetric Gauge Theories, hep-
th/9803026.
10. M. Green, J. Schwarz and E. Witten, Superstring Theory, Vols I and II, Cambridge
University Press, 1987.
11. D. Liist and S. Theisen, Lectures in String Theory, Lecture Notes in Physics, 346,
Springer Verlag, 1989.
12. L. Alvarez-Gaume and M. Vazquez-Mozo, Topics in String Theory and Quantum
Gravity, in the 1992 Les Houches School, Session LVII, eds. B. Julia and J. Zinn-
Justin, Elsevier Science Publishers, 1995.
13. J . Polchinski, "What is String Theory?", hep-th/9411028.
14. H. Ooguri and Z. Yin, hep-th/9612254.
15. E. Kiritsis, "Introduction to superstring theory," Leuven University Press, 1998;
hep-th/9709062.
16. J. Polchinski, "String Theory", Cambridge University Press,1998.
17. M. Duff, R. Khuri and J. Lu, " String Solitons", Phys. Rept. 259 (1995) 213,
hep-th/9412184.
18. J. Polchinski, S. Chaudhuri and C. Johnson, "Notes on D-branes", hep-th/9602052.
19. J. Maldacena, "Black Holes in String Theory", PhD Thesis; hep-th/9607235.
20. J. Polchinski, "TASI lectures on D-branes," hep-th/9611050.
21. S. Forste and J. Louis, "Duality in String Theory", hep-th/9612192.
22. C. Vafa, "Lectures on Strings and Dualities", hep-th/9702201.
23. E. Kiritsis, "Introduction to Non-Perturbative String Theory", hep-th/9708130.
24. W. Lerche, "Recent Developments in String Theory", hep-th/9710246.
25. A. Peet, " The Bekenstein Formula and String Theory (N-brane Theory)", hep-
th/9712253.
26. B. de Wit and J. Louis, "Supersymmetry and Dualities in Various Dimensions",
hep-th/9801132.
27. W. Taylor, "Lectures on D-branes, Gauge Theory and M{atrices)", hep-
th/9801182.
28. A. Sen, " An Introduction to Non-Perturbative String Theory", hep-th/9802051.
29. C.P. Bachas, "Lectures on D-branes" hep-th/9806199.
30. O. Aharony, S.S. Gubser, J. Maldacena, H. Ooguri and Y. Oz, hep-th/9905111.
31. G.'t Hooft, Nuc!. Phys. B19 (1974) 276.
32. A. Polyakov, JETP Lett. 20 (1974) 194.
33. B. Julia and A. Zee, Phys. Rev. Dll (1975) 2227.
34. P. Goddard, J. Nuyts and D. Olive Nucl. Phys. B125 (1977) 1.
35. C. Montonen and D. Olive, Phys. Lett. B12 (1977) 117.
36. E. Witten, Phys. Lett. B86 (1979) 283.
37. J. Bagger and J. Wess, "Supersymmetry and Supergravity" Princeton Series in
Physics, second edition, 1992.
38. R. Haag, J. Lopuszanski and M. Sohnius, Nucl. Phys. B88 (1975) 257.
39. A. Sen, Phys. Lett. B329 (1994) 217; hep-th/9402032.
40. G. Segal and A. Selby, Commun. Math. Phys. 177 (1996) 775.
41. C. Vafa and E. Witten, Nucl. Phys. B431 (1994) 3 hep-th/9408074.

42. N. Seiberg and E. Witten, Nuc!. Phys. B426 (1994) 19; hep-th/9407087
Nucl. Phys. B431 (1994) 484 hep-th/9408099.
43. S. Mandelstam, Phys. Lett. B53 (1975) 476; Phys. Rev. D19 (1979) 2391.
44. G. 't Hooft, NucI. Phys. B190 (1981) 455.
45. A. Klemm, W. Lerche, S. Yankielowicz and S. Theisen, Phys. Lett. B344 (1995)
169 hep-th/9411048.
46. P.C. Argyres and A.E. Faraggi, Phys. Rev. Lett. 74 (1995) 3931 hep-th/9411057.
47. N. Seiberg and E. Witten, Nuc!. Phys. B426 (1994) 19 hep-th/9407087.
48. L. Alvarez-Gaume, J. Distler, C. Kounnas and M. Marino, Int. J. Mod. Phys. All
(1996) 4745; hep-th/9604004.
49. G. Veneziano, Nuovo Cimento, 57A (1968) 190.
50. J. Scherk and J .H. Schwarz, Nucl. Phys. B81 (1974) 118.
51. M.B. Green and J.H. Schwarz, Phys. Lett. 149B (1984) 117.
52. L. Alvarez-Gaume and E. Witten, Nucl. Phys. B234 (1984) 269.
53. 1. Antoniadis, Phys. Lett. B246 (1990) 377.
54. N. Arkani-Hamed, S. Dimopoulos and G. Dvali, Phys. Lett. B429 (1998) 263;
hep-ph/9803315.
55. 1. Antoniadis, N. Arkani-Hamed, S. Dimopoulos and G. Dvali, Phys. Lett. B436
(1998) 257; hep-ph/9804398.
56. N. Arkani-Hamed, S. Dimopoulos and G. Dvali, Phys. Rev. D59 (1999) 086004;
hep-ph/9807344.
57. E. Kiritsis and C. Kounnas, Phys. Lett. B331 (1994) 51 hep-th/9404092.
58. E. Kiritsis and C. Kounnas, gr-qc/9509017.
59. D.J. Gross, J.A. Harvey, E. Martinec and R. Rohm, Nuc!. Phys. B256 (1985) 253;
Nuc!. Phys. B267 (1986) 75.
60. C. Hull and P. Townsend, Nucl. Phys. B438 (1995) 109; hep-th/9410167.
61. E. Witten, Nucl. Phys. B443 (1995) 85 hep-th/9503124; hep-th/9507121.
62. R. Nepomechie, Phys. Rev. D31 (1985) 1921.
63. C. Teitelboim, Phys. Lett. B167 (1986) 69.
64. S.R. Das, Nucl. Phys. Proc. Supp!. 68 (1998) 119 hep-th/9709206.
65. E. Kiritsis, JHEP 10 (1999) 010 hep-th/9906206.
66. C.G. Callan and J.M. Maldacena, Nucl. Phys. B472 (1996) 591 hep-th/9602043.
67. S.R. Das and S.D. Mathur, Nuc!. Phys. B478 (1996) 561 hep-th/9606185.
68. J. Maldacena and A. Strominger, Phys. Rev. D55 (1997) 861 hep-th/9609026.
69. G. 't Hooft, Nuc!. Phys. B72 (1974) 46l.
70. A.M. Polyakov, Nuc!. Phys. Proc. Supp\. 68 (1998) 1 hep-th/9711002.
71. G.T. Horowitz and A. Strominger, NucI. Phys. B360 (1991) 197.
72. J. Maldacena, Adv. Theor. Math. Phys. 2 (1998) 231 hep-th/9711200.
73. S.S. Gubser, I.R. Klebanov and A.M. Polyakov, Phys. Lett. B428 (1998) 105 hep-
th/9802109.
74. E. Witten, Adv. Theor. Math. Phys. 2 (1998) 253 hep-th/9802150.
75. J . Maldacena, Phys. Rev . Lett. 80 (1998) 4859 hep-th/9803002.
76. S. Rey and J. Yee, hep-th/9803001.
77. C. Csaki, H. Ooguri, Y. Oz and J. Terning, JHEP 01 (1999) 017 hep-th/9806021.
78. E. Witten, Adv. Theor. Math. Phys. 2 (1998) 505 hep-th/9803131.
79. I.R. Klebanov and A.A. Tseytlin, Nucl. Phys. B546 (1999) 155 hep-th/9811035.
BARYOGENESIS
MIKHAIL SHAPOSHNIKOV
Institute of Theoretical Physics
Lausanne University, BSP-Dorigny
CH-l015 Lausanne, Switzerland
1. Introduction
Why does the Earth, Moon, Sun, planets, our galaxy, including ourselves,
consist of matter (i.e. protons, neutrons and electrons), but not of antimat-
ter? In addition, Why do we not see any considerable amounts of antimatter
(i.e. antiprotons, antineutrons and positrons) in the Universe? These are the
questions we address when we talk about the problem of baryon asymmetry
of the Universe.
Before 1932, the problem could not be formulated at all, since the only
elementary particles then known were the proton, the neutron and the
electron. Antimatter was unknown - there was nothing to discuss.
The birth of antimatter can be related to the theoretical work of P.
Dirac. His relativistic equations, describing spin-lj2 particles, predicted the
existence ofthe positively-charged electron (the positron). This particle was
discovered experimentally soon after. In his Nobel lecture in 1933, Dirac
said [1]: "If we accept the view of complete symmetry between positive and
negative electric charge so far as concerns the fundamental laws of Nature,
we must regard it rather as an accident that the Earth (and presumably
the whole solar system) contains a predominance of electrons and positive
protons. It is quite possible that for some of the stars it is the other way
about, these stars being built up mainly of positrons and negative protons.
In fact, there may be half the stars of each kind. The two kinds of stars
would both show exactly the same spectra, and there would be no way
of distinguishing them by present astronomical methods." Thus, right after
the discovery of antimatter a new picture of the Universe emerged, in which
it would be symmetric with respect to matter and antimatter.
Present astronomical observations do not support Dirac's hypothesis
that the Universe contains matter and antimatter in equal amounts. An-
397
1.-1. Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments. 397-416.
398 MIKHAIL SHAPOSHNIKOV
timatter in the Universe can be detected by two different means. First, if

antigalaxies exist, we should see antiprotons and antinuclei in cosmic rays,
precisely in the same way as we see cosmic protons and nuclei. However, no
antinuclei have been observed in cosmic rays. The positrons, antiprotons
and antineutrons are observed. These antiparticles can be produced in col-
lisions of cosmic protons or nuclei with galactic gas and with particles in
the Earth's atmosphere, and their amount is consistent with expectations
[2].
Second, in the regions where matter and antimatter are mixed, an-
nihilation of protons and antiprotons must take place. Annihilation will
produce about 5-6 71"0 and 7I"± mesons which, in turn, will decay into ,-
quanta, electrons, positrons, neutrinos and antineutrinos. The spectrum
of annihilation ,-quanta has a specific form, with energy peaked around
(2GeV)/(5 - 6)/2 '" 150 MeV. However, this has not been observed [2].
Therefore, we arrive at the conclusion that the observable Universe, with
size R '" 3000 Mps '" 1010 light-years, is globally asymmetric and contains
no antimatter. So, the baryon asymmetry of the Universe (BAU) is a real
fact, which requires an explanation. A safer, but very unlikely [3], statement
is: the Universe consists of islands of matter and antimatter of the size of
clusters of galaxies, f '" 20 Mps, containing around 3000 galaxies. These
islands must be separated by empty space, to prevent pp annihilation. Even
in the latter, symmetric case, one should be able to understand why the
Universe is asymmetric on cosmological distances and how this complicated
structure with islands of matter and antimatter separated by voids could
emerge.
From Dirac's discovery until about 1956, it was believed that the laws
of Nature are completely symmetric with respect to the charge conjuga-
tion (change of particle to antiparticle). Thus, baryon asymmetry is very
strange from this point of view. The situation changed slightly in 1956,
when the discovery of the breaking of C and P parities was made. Thus,
some properties of particles and antiparticles can be different. Neverthe-
less, this fact does not help in understanding of the problem of the baryon
asymmetry of the Universe. One can construct a combination of C and P
symmetries, namely CP parity, which changes a particle to an antiparticle
and simultaneously changes the momentum of the particle. This symmetry
was believed to be exact at that time, so the breaking of C sheds no light
on our main problem.
The breakthrough happened in 1964 and 1965, with two major discover-
ies. The first one dealt with particle physics and the second with cosmology.
In 1964, CP non-conservation was discovered [4] in decays of KO mesons.
Thus, there is a tiny difference (experimentally, about 10-3 ) between the
properties of matter and antimatter. Still, it is completely unclear how
BARYOGENESIS 399
this extremely small difference could produce the 100% baryon asymmetry
we observe in the Universe, (nB - nfJ)/(nB + nS) = 1 (nB and nfJ are
concentrations of baryons and antibaryons correspondingly).
The second discovery, from the cosmology side, helps a lot in under-
standing the latter issue. In 1965 it was found [5] that the Universe is filled
by cosmic microwave background radiation (CMBR) - thermal photons
with a Planck spectrum and a temperature of 2.73K. This was an experi-
mental confirmation of the Big Bang theory suggested by George Gamow
in 1946 [6]. Let us first look at the content of our Universe at the present
moment. In average, it has about 400 relic photons per cubic centimetre,
approximately the same number of relic neutrinos, and about 10- 7 nucle-
ons. The present photon density is related to the temperature of CMBR as
n-y = 0.244T 3, while the density of nucleons can be very roughly estimated
from the amount of visible matter in the Universe. A more refined estimate
for 'TJ == nB/n-y = (1.5 - 6.3) x 10- 10 follows from the comparison of the
light element abundances with the predictions of Big Bang nucleosynthesis,
see, e.g. [7].
The Universe expands and now has thermal photons. Thus in the past it
was very hot and dense. An interesting reference point for us in the history
of the early Universe is at t = 10- 6 s from the beginning of the Big Bang. At
this point the temperature of the Universe was about 1 GeV. At this time
the number densities of particles were huge, n-y ~ nB rv 1040 cm- 3 . Since
the plasma was hot, with a temperature higher than the masses of light
quarks, the number of quarks and antiquarks is the same as the number
of photons, up to spin factors. Thus, for each 1,000,000,000 photons, we
will find approximately 1,000,000,000 quarks and 999,999,999 antiquarks.
When the Universe cools down from this state, what happens is that these
999,999,999 baryons and antibaryons annihilate into photons and neutrinos,
but one nucleon does not find a pair and survives. The ratio of nucleons and
photons after annihilation is nearly equal to the number we observe today,
namely nB/n-y ~ (1.5 - 6.3) x 10- 10 . The nucleon that survives gives rise
to galaxies, stars and planets.
Thus, the hot Big Bang theory requires the baryon asymmetry of the
Universe to be a tiny number at the initial stages of the Universe's expan-
sion, (nB - nfJ)/(nB + nn) ~ 10- 9 - 10- 10 . Interestingly, this number is
much smaller than the "quantitative measure" € ~ 10- 3 of the difference
between matter and antimatter, observed in K O decays.
In 1967 Andrei Sakharov suggested [8] that the Universe is asymmet-
ric because the baryon number is in fact non-conserved. In this case, the
Universe could start its expansion from a truly symmetric state, containing
an equal number of particles and antiparticles. Then, in the course of the
expansion, the particle physics reactions with Band C P non-conservation
would produce an excess of particles over antiparticles. He assumed that

there exist some heavy particles, with masses of the order of the Planck
scale, 1019 GeV, which can decay with baryon number non-conservation and
C P violation. For example, one can take "leptoquarks" X of grand unified
theories, with the following modes of decay: X -+ qt, iiii and X -+ iiZ, qq
(q is quark and 1 is lepton). Now, if C P is broken, an equal number of X
and X will, after their decay, leave a different number of quarks and anti-
quarks, precisely as the decays of KO and [(0 mesons leave different numbers
of electrons and positrons. It is sufficient to produce a small asymmetry,
rv 10- 9 - 10- 10 , which will then be converted into a 100% asymmetry after
the annihilation of matter and antimatter.

The model for baryogenesis suggested by Sakharov (see also the some-
what later paper by Kuzmin [9]) was, perhaps, too naive, and we still do
not know for sure how the baryon asymmetry has emerged. This problem
has attracted a lot of research, and hundreds of papers have been published
on this topic. The problem is situated at the border of different fields. From
particle physics it requires Band C P violation. Therefore, it provides hints
for model building, puts constraints on the amplitudes of different reactions,
and provides motivation for the experimental search for proton decay. All
these processes occur in the high-temperature plasma. Thus, it is a natural
application of the methods of equilibrium and out-of-equilibrium statistical
physics. Finally, it deals with cosmology and astrophysics, putting different
constraints on possible scenarios of the early Universe evolution.
In this lectures I will consider different mechanisms for baryogenesis.
Most part of this contribution to the proceedings is plagiarism from the
paper I wrote for Contemporary physics [10]. The reader interested to learn
more about anomalous B-non-conservation at high energies and tempera-
tures, phase transitions in the early Universe and electroweak baryogenesis
may find many details in ref. [11], together with references to original pa-
pers.
2. General conditions for baryogenesis
Before going to different mechanisms for baryogenesis, let us try to under-

stand the general conditions needed for appearance of baryon asymmetry.
In fact, we know very little about the state of the very early Universe.
The observation of the cosmic microwave background radiation (CMBR)
with thermal spectrum tells us that the Universe was hot with T rv 0.3 eV
at the time of rv 3 X 1013 s from the beginning of the Big Bang (at this time
recombination occurs, freely moving electrons and protons form neutral
atoms, and the Universe starts to be transparent to radiation). The success
of the theory of the Big Bang nucleosynthesis tells that the Universe was
BARYOGENESIS 401
hot even earlier, at temperatures of the order of 0.1 MeV. Any hypotheses
about the state of the Universe at earlier stages are based on extrapolation.
Nevertheless, a number of theoretical considerations and observations sug-
gest a quite consistent picture. This picture looks as follows. At the very
beginning, the Universe was in a state with energy density of the order of
the Planck scale Mf,l' Mpi = 1.22 X 1019 GeV. The baryon number density
of this state is, in fact, irrelevant. After some time, inflation, which is a
period of the exponential expansion of the Universe, starts. During infla-
tion, any baryon density that was present before, is diluted exponentially.
The inflation is driven by some so-called inflaton field. At some time this
field decays and produces the entropy of the Universe. After the inflaton
decay the Universe is radiation-dominated. So it is natural to expect that
at the end of inflation the baryonic density is essentially zero, and baryon
asymmetry should thus be created afterwards.
The mechanism of baryon asymmetry creation is related to the par-
ticle physics model. Several general remarks are applicable. First, baryon
number should not be conserved. Otherwise, the state, initially symmet-
ric with respect to the baryon number, cannot evolve into the state with
non-zero baryon number we observe today. Another requirement is that the
theory should distinguish between particles and antiparticles. This means
that the C and CP symmetries should be broken. We will consider these
symmetries further later. The third requirement is the absence of thermal
equilibrium (Sakharov called this the "arrow of time"). Let us discuss this
last requirement in more detail.
Assume for a moment that the system is indeed in thermal equilibrium.
This is a very strong statement. It allows an immediate definition of all the
properties of the plasma and their expression by very few parameters, such
as the temperature and the densities of conserved charges. The existence
of any conserved charge is quite unlikely at the Planck scale, and thus
the thermal equilibrium state is characterized by only one parameter -
the temperature. If some charge is non-conserved (and we want baryon
charge to be non-conserved), then in thermal equilibrium it is equal to
zero, independently of the conservation or non-conservation of C and CPo
To be even more clear: thermal equilibrium is a state in which there is no
time evolution (by definition). Thus, if baryon charge was equal to zero at
the beginning, it will be zero all the time.
Of course, the Universe expands, and perfect thermal equilibrium is not
possible. Qualitatively, deviations from thermal equilibrium are given by
the ratio of the reaction rates and the rate of the Universe expansion. If
the theory that describes the particle interactions is known, the equation of
state (the relation between energy density and pressure) can be readily com-
puted. The equation of state determines the rate of the Universe expansion.
Moreover, one can compute the rates of different particle physics reactions
in the system, assuming, for example, that the particle densities are ther-
mally distributed. Roughly, a particular reaction is in thermal equilibrium
if its rate greatly exceeds the rate of the Universe expansion. If true, the
distribution of the particles follows thermodynamics, and deviations from
equilibrium are small.
Let us take, for example, electromagnetic processes, such as e+ e- ~
'Y'Y. The typical time of reaction is given by T (unv)-l, where u is the
I'V
annihilation cross-section of the reaction(roughly speaking, the size of the

electron), n is the concentration of electrons, and v is a relative velocity. At
high temperatures T > me, n T3, V 1 (in units of the velocity of light)
I'V I'V
and the cross-section is of the order of u I'V0:2 jT2, where 0: is the fine-
structure constant (I'V1/137). We can compare this with the time of the
Universe expansion, t MoIT2 (where Mo = Mpz/1.66v'N and N
I'V I'V100 is
the number of effectively massless degrees offreedom) and discover that this
reaction is in thermal equilibrium at temperatures below T ~ 0:2 Mo 1014 I'V
GeV. It goes out of thermal equilibrium at temperatures smaller than the

electron mass, since the concentration of electrons exponentially goes to
zero, n I'V exp( -melT). So, deviations from thermal equilibrium for this
reaction are very small for quite a long time, starting almost from the
Planck scale and down to the electron mass.
The requirement to be out of thermal equilibrium is thus very im-
portant. Imagine that there are some reactions with baryon number non-
conservation. If they are in thermal equilibrium, baryon asymmetry cannot
be generated at this time; moreover, any excess of baryon number that ex-
isted before this epoch will be essentially diluted to zero. Baryogenesis must
therefore take place when reactions with baryon number non-conservation
are out of thermal equilibrium. This is somehow a paradoxical requirement:
reactions are out of equilibrium when they are slow enough (slower than the
expansion rate of the Universe). However, if they are slow, perhaps they do
not generate much asymmetry. This gives a quantitative requirement: the
amount of baryon asymmetry produced does depend on the rates although
the fact of the asymmetry generation does not. The summary of this dis-
cussion is usually dubbed as the three Sakharov necessary conditions for
baryogenesis: baryon number non-conservation, C and CP violation, and
breaking of thermal equilibrium. Now we are going to see how these condi-
tions are satisfied in different particle physics models.
3. Anomalous electroweak fermion number non-conservation
On a perturbative level the electroweak theory has four conserved fermionic

numbers: B - baryon number, and L e, LJ.I' Lr - leptonic numbers. Quantum
BARYOGENESIS 403
anomaly
8,Jff = 8,Jf; = 3~;2 Tr (F'JV FI'II) + U(I) part
leads to anomalous processes with non-conservation of baryon and lepton
number:
bosons +-+ bosons + 9q + 31.
Here JB and JL are baryon and lepton currents, F",v is the SU(2) field
strength, nf is the number of fermion generations, q and 1 are quarks and
leptons.
1
As a result only 3 numbers are conserved: Li - B. Protons are stable
but deuterium can decay into antineutron, positron and antineutrinos:
D = pn -7 n e+fil'fiT .
The rate of anomalous fermion-number non-conservation at zero and non-
zero temperatures is of the order of (see recent papers [12, 13] and reviews
[11, 14]):
exp( - ~:, ) I"V 10- 160 , T=O
rl"V (1)
where M sph Mw / ow is the sphaleron mass and Tc is the temperature of

I"V
the electroweak (EW) phase transition. Further and more detailed discus-
sion can be found in refs. [11, 15].
4. Electroweak baryogenesis
Let us consider the quality of thermal equilibrium at the electroweak scale,
where baryon-number violating processes may potentially lead to baryon
asymmetry.
The measure of deviation from the thermal equilibrium is the ratio of
two time scales. The first one is the rate of the Universe expansion, given by
the inverse age of the Universe t r/: tu = ~. Here Mo = Mpz/1.66N~ I"V
1018 Ge V, and N is the effective number of the massless degrees of freedom.

The expansion rate of the Universe is a unique non-equilibrium parameter
of the system (at least in the absence of different phase transitions). The
second time scale (different for different types of interaction) is a typical
reaction time, given by (Treaction)-l (O'nv), where 0' is the corresponding
I"V
cross-section, n is the particle concentration and v is the relative velocity

of the colliding particles.
The fastest reactions at temperatures T rv mw are those associated with

strong interactions (e.g. qij -+ GG); their rate is of the order of (Ts trong)-1 rv
a;T. The typical weak reactions, say ev -+ ev, occur at the rate (Tweak) -1 rv
a~ T, and the slowest reactions are those involving chirality flips for the
lightest fermions, e.g. eRH -+ vW with the rate (Te )-1 rv {;awT, where
ft
Ie is the electron Yukawa coupling constant. Now, the ratio varies from
10- 14 for the fastest reactions to 10-2 for the slowest ones. This means that
particle distribution functions of quarks and gluons, intermediate vector
bosons, Higgs particle and left-handed charged leptons and neutrino are
equal to the equilibrium ones with an accuracy better than 10- 13 ; the
largest deviation from thermal equilibrium (rv 10- 2 ) is being expected for
the right-handed electron.
We see that the Universe at the electroweak scale is very close to ther-
mal equilibrium. How then we could baryon asymmetry at all? A way out
is provided by a first order electroweak phase transition. At T > Tc the
symmetry is restored, and the vacuum expectation value of the Higgs field
is close to zero [16, 17](see Fig.1).
At T = Tc the free energy of the broken phase is equal to the free en-
ergy of the restored phase. However, the phase transition does not occur at
this time because the probability of tunnelling through the barrier is zero.
So, it happens somewhat later and goes through the bubble nucleation. at
the electroweak scale the size of the event horizon is := 1 cm. The critical
bubbles are of microscopic size, roughly R rv (awT)-1. Bubbles nucleate
in different places, expand, and eventually collide, filling the Universe with
a new phase. The typical size of bubbles at the time they collide is macro-
scopic, rv 10-6 cm. Inside the bubble the vacuum expectation value of the
Higgs field is non-zero, and it is zero outside it. So, inside the bubble the
rate of B-violation is small, whereas outside the bubble the rate is large.
The masses of particles are different inside and outside the bubbles (parti-
cles inside the bubble are generically heavier, since their masses arise from
interactions with a Higgs field).
The electroweak phase transition, if it is strongly first order, is quite
a violent event, during which concentrations of particles may be far from
thermal equilibrium because of the motion of the bubble walls. This gives
rise to a strong violation of thermal equilibrium, which is a necessary con-
dition for successful baryogenesis.
Let us see what happens when the domain wall propagates through the
plasma. Take an expanding spherical bubble filled by the Higgs phase (see
Fig. 2). Suppose that a domain wall moves with velocity v (typically, v may
vary from 10- 3 to 0.1). Let us pick up a small portion on the domain wall
(bottom part of the bubble on Fig. 2) and take a frame associated with it.
We will feel a "wind" of particles, moving from the bottom to the top. The
BARYO GENESIS 405
0.5 V( lj»
0.25
-0.25
-0.5
-0.75
-1 U-______ __ ______ _____ _ w

50 100 150
Figure 1. The evolution of the effective potential with temperature.
velocities of individual particles are rather large, of the order of the speed of
light, but the velocity of their organized motion (the velocity of the domain
wall) is much smaller. Particles may go through the domain wall if their
energy allows them to do so, or may be reflected. On top of that, particles
interact with other particles. Let us take, for instance, a particle that was
reflected, and which moves down after the reflection. Since there are other
particles, approaching the domain wall from the bottom, with which the
reflected particle interacts, it will be stopped after some time; it will then
join other particles moving towards the domain wall and eventually pass
through it. Thus, the concentration of particles in front of the domain wall
is larger than that far from it. The distance from the domain wall where
the density disturbance disappears is determined by a mean free path A of
particles (this is the distance a particle travels without collisions) in the
plasma and by the velocity of the domain wall. The smaller the velocity
of the wall is and the larger the mean free path is the larger will be the
distance the reflected particle can travel out of the domain wall. The typical
size of the disturbed region (region with an excessive density) is of the order
antiquark
incident
qual1\ antiqual1\
SYMMETRIC PHASE
B Isnol conserved
S is 2;ero
Figure 2. The expanding bubble. The dark region inside the small circle corresponds to
the broken (Higgs) phase where quarks and leptons are massive and baryon number vio-
lating processes are strongly suppressed. The region between the two circles corresponds
to the plasma in the symmetric phase which is disturbed by the motion of the bubble
wall. It is this region which is responsible for the generation of baryon excess. The rate
of baryon number non-conservation is high here, while OP-non-invariant interaction of
fermions with the domain wall spatially separates fermions from antifermions in a way
that fermions are going inside the bubble. The region outside the large circle corresponds
to the undisturbed symmetric phase.
of [I'V >../v, and the typical time the particle reflected from the domain wall
spends in the symmetric phase before it is absorbed by it is t I'V >"/ v 2 • The
magnitude of the density perturbation must be proportional to the domain
wall velocity, just because if v = 0 we have thermal equilibrium. It can
be concluded from this that deviations from thermal equilibrium at the
EW phase transition may be large, and are in fact not directly connected
with the extremely slow (in comparison with the rates of particle physics
reactions) Universe expansion.
CP violation means that particles and antiparticles interact with the
bubble wall in a different way. Let us take an idealized situation, in which
the domain wall is at rest and there is no thermal bath, and consider what
BARYOGENESIS 407
happens if a quark moves towards the domain wall. The quark may be
reflected with a probability R, or may be transmitted with a probability
T; the total probability of interaction is thus equal to 1, R + T = 1. The
corresponding probabilities for the antiquark are denoted by 1', fl, and we
have 'i' + fl = 1. Because of CP violation, R i- fl, and consequently, T i- 'i'.
Now, consider a thermal distribution of particles, but neglect their interac-
tions with one another. Take first particles and antiparticles that approach
the domain wall from the symmetric phase (we again have in mind the
bottom part of the domain wall in Fig. 2). If equal numbers of quarks and
antiquarks, with the same velocity distribution, move towards the domain
wall, the numbers of reflected quarks and antiquarks will not be equal. To
have a complete picture, we should take into account that there are quarks
and antiquarks that move towards the domain wall from inside the bubble.
Because of CP violation, the transmission and reflection probabilities are
different for quarks and antiquarks, T i- 1", R i- Il'. The CPT theorem
says, however, that T' + R = 'i" + Il, so that the total number of particles
moving up is equal to the corresponding number of antiparticles. This is as
it should be in thermal equilibrium. When the domain wall moves through
the plasma, the number of particles and antiparticles approaching the do-
main wall from the bottom is larger than that from the top. So the total
number of particles that were incident in the symmetric and the broken
phases, and which find themselves in the symmetric phase after interacting
with the domain wall, is no longer equal to the number of antiparticles [18].
To put it in another way: CP-non-invariant interactions of fermions with
moving domain wall create a net baryon current, which is directed inside
the bubble, its sign depending on the sign of CP violation. Assume that
antibaryons move into the symmetric phase, and baryons to the broken
one. Basically, the baryon number is conserved in the broken phase and
not conserved in the symmetric one. The antiquark, as we discussed above,
is exposed to the symmetric phase for a time to rv A/V. lithe rate offermion
number non-conservation is large enough, the excess of antiquarks will dis-
appear, just because B-non-conservation is in thermal equilibrium in the
symmetric phase. li the time to is small, then only part of the antifermion
excess is diluted, and the total flux of baryons, which enters in the bro-
ken phase, is decreased. So the baryon number density inside the bubble is
non-zero, the region of the broken phase increases and, at the time different
bubbles collide, the whole Universe is baryon-asymmetric [18, 19, 20].
The above discussion is oversimplified and does not take into account
many effects that are essential for a quantitative analysis. For example,
if the width of the domain wall is larger than or comparable with the
mean free path of the particles (which looks like a realistic case), it is clear
that the simple quantum-mechanical transmission-reflection picture must
symmetric phase
critical point
Higgs phase
Figure 3. The phase diagram of the electroweak theory. For small Higgs masses, to
the left from the critical point, the electroweak phase transition is of the first kind. At
the critical point it is of the second kind, and the critical properties of the electroweak
theory near this point are very similar to those of the liquid-vapour system. At large
Higgs masses, to the right from the critical point, there is no phase transition at all.
be modified [20]. In this case particles in the plasma interact with one
another, so it is better to talk about quasiparticles than particles [21], etc.
The discussion of such details goes beyond the scope of this paper (for a
recent review see [11] and references therein, and also [22, 23]).
In order to see if the baryon asymmetry of the Universe arises because of
the electroweak anomaly, explicit computations should be performed within
some model. The MSM is an obvious candidate. The first condition to check
is whether there is a first-order phase transition which is strong enough to
suppress baryon non-conservation right after the EW phase transition [24].
The phase diagram for the electroweak theory is shown in Fig. 3 [25].
The vertical axis is the temperature, while the horizontal axis is the Higgs
mass in the MSM. In many extensions of the standard model the phase
diagrams look qualitatively the same, but in all of these the horizontal axis
is a combination of the different parameters rather than the Higgs mass
itself. This diagram is similar to that of a liquid-vapor system. There is an
end-point of a line of the first-order phase transitions. If the Higgs mass is
equal to the critical value, the phase transition in the system is of second
order. For smaller Higgs masses the EW phase transition is of the first kind,
BARYOGENESIS 409
while at larger Higgs masses there is no phase transition at all.

For a minimal standard model, where all parameters besides the Higgs
mass are fixed by experiment, the end point of the first-order phase tran-
sition line can be computed unambiguously and corresponds to MH ::: 72
GeV [26]. If the minimal standard model is right, then the direct LEP ex-
perimental bound on the Higgs mass is mH > 95.2 GeV [27]. Thus, in the
realistic MSM there is no phase transition at all so the minimal standard
model can be excluded as a theory for electroweak baryogenesis. Exten-
sions of the standard model may be viable from the point of view of the
strength of the phase transition. The minimal standard model contains just
one scalar particle - the Higgs boson. Extra scalars in the mass range rv 100
Ge V change the strength of the phase transition. For example, in a theory
with two Higgs doublets or in the minimal supersymmetric standard model,
the phase transition may be strong enough to prevent the baryon asym-
metry from sphaleron erasure. There can be a very strong first-order phase
transition, for example, if the lightest Higgs mass is somewhere between 75
and 105 GeV and the right scalar top mass mR between 100 GeV and 160
GeV [28]. This set of parameters is constrained by existing experiments,
and is achievable in future experiments at the electron-positron collider at
CERN [29].
Assuming now that we have a model with the necessary first-order phase
transition, what is the magnitude of the baryon asymmetry created? Clearly,
it depends on the particle content, on the strength of CP-violation, on the
rate of B-non-conservation in the symmetric phase, on the bubble wall ve-
locity, and on many other details of interactions. Computations of baryon
excess have been carried out in different models, and results are quite en-
couraging: the number 10- 10 may be derived (look at almost any of the
papers in the list of references), provided the parameters of the model are
chosen in some particular way. In addition to constraints on the particle
spectrum, following from the requirement of a first-order phase transition
(which may be quite demanding [28, 29]), there appear to be extra con-
straints on the model, since a sufficient amount of CP-violation is required.
In the MSM with three fermion generations the only source of CP-
violation is the Kobayashi-Maskawa phase. Because of the quite specific
structure of Kobayashi-Maskawa CP-violation it is unlikely that it can be
responsible for the observed baryon asymmetry (different mechanisms for
the amplification of MSM CP-violating effects were suggested in [24, 21,
30]). Naturally, extensions of the standard model do contain extra sources
of CP-violation. This makes baryogenesis easier; moreover, computing the
baryon excess and comparing it with observations gives some idea of the
magnitude of different CP-violating phases, specific for some MSM exten-
sion. This is very interesting, since new sources of CP-violation may reveal
themselves in other situations, rather than KO decays only, which may

be looked for experimentally. One of the CP-violating effects which is ex-
tremely small in the MSM and is far beyond experimental reach is the
neutron electric dipole moment (it must be zero if CP is conserved). In dif-
ferent extensions of the standard model the dipole moment is much larger
and is comparable to the experimental limit 1O- 23 e cm, where e is the elec-
tron charge. If the baryon asymmetry of the Universe has an electroweak
origin then it is likely that the neutron dipole electric moment is large
enough to be observed experimentally and that new CP-violation can be
found beyond KO physics, e.g. in B (beauty) meson decays. Unfortunately,
constraints on CP-violating phases from the requirement of EW baryogen-
esis are quite uncertain at the moment. For example, in the MSSM it is the
CP-violation in the chargino sector that plays the most important role for
baryogenesis. According to various authors, the lower limit for the phase
o in the chargino mixing matrix should be 0.1, 0.01, 0.001 or even 0.0001
(values given in [22, 31, 23, 32]).
5. GUT baryogenesis
The source of non-conservation of baryon number in GUT baryogenesis
is associated with unification of strong, weak and electromagnetic interac-
tions. A mechanism was already suggested in Sakharov's paper of 1967, and
then elaborated by Kuzmin in 1970 [9], Ignatiev, Krasnikov, Kuzmin and
Tavkhelidze in 1977 [33] and Weinberg in 1979 [34]. In ref. [8] it was assumed
that there exist some superheavy particles (they were called "maximons" in
the original paper, but we will use a modern terminology - "leptoquarks"),
which decay with baryon number non-conservation and CP-violation. For
example, leptoquarks X of grand unified theories can decay as X -+ qt, ijij
and X -+ ijl, qq. If CP is broken, an equal number of X and X will, af-
ter their decay, leave a different number of quarks and antiquarks. If the
universe were as hot as the leptoquark mass Mx and in a state close to
thermal equilibrium then baryon asymmetry of the Universe resulting from
leptoquark decays is of the order
1
l:!.. '" -N. oCP . Smacro, (2)
ell
where oCP is the asymmetry in leptoquark decays,
o _ r(X -+ qq) - r(X -+ ijij)
(3)
CP- r tot '
r tot is the total width, and Smacro is a factor taking into account the kinetics
M2
of the leptoquark decays: Smacro '" liPI
T.tot for Mx
2
< TtotMpl and Smacro '"
BARYOGENESIS 411
1 for the opposite case. Here Mpl '" 1019 GeV is the Planck mass and Neff
is the number of massless degrees of freedom.
Under the assumption that the universe had temperatures of the req-
uisite order of magnitude, many grand unified theories give rise to baryon
asymmetry. Some care should be taken on the equilibrium character of
anomalous electroweak reactions with B-non-conservation. Several differ-
ent cases can be distinguished, depending on initial conditions and on the
rate of B- and L- non-conservation due to processes other than those asso-
ciated with sphalerons.
(i) Supposing that the Universe is asymmetric with respect to the anomaly-
free fermionic charges Ll i = Li - .l..
nf
B of the standard model at T > T**,
and assuming that at T < T** there are no B- or L- violating interactions
besides the electroweak anomalous processes (the origin of the primordial
asymmetry is not essential here) then anomalous reactions convert the ini-
tial asymmetry to the baryonic one at T = T*. For the minimal standard
model the relationship is given by [35J
nf m?(T*)
Ll = O(l)LlB-L - 0(10- 2 ) ~ (~*)2 Lli' (4)
where m~ is the lepton mass of a given generation. The first term on the
right-hand side ofthis equation tells that the B-L asymmetry is reprocessed
into the baryon asymmetry; the second term is the correction coming from
the slightly different behavior of quarks with different masses in the plasma.
If the initial value of B-L is non-zero (from the GUT physics, say) then the
baryon asymmetry, apart from a possible contribution from the EW phase
transition (see below), has a primordial character. If, on the contrary, the
initial B-L asymmetry is absent, we can rely only on the second term in
(4). For three lepton generations there is a suppression Llo ~ 3 X 1O- 6 Ll 3 •
So, to have a non-negligible effect, the initial asymmetry Ll3 must be very
large, or the standard theory should be extended by adding heavy leptons.
(ii) Suppose now that there are some reactions which do not conserve
all Ll i and which are in thermal equilibrium for some period between Tc and
T**. At this intermediate period Band L are non-conserved separately, and
according to the non-equilibrium Sakharov condition all baryonic and lep-
tonic asymmetries are washed out. Hence, the existence of these reactions
is fatal for the primordial baryon asymmetry. IT the baryon asymmetry is
not produced at a later time, the requirement of the absence of these reac-
tions may appear to be a powerful tool for constraining the properties of
new particle interactions [36t[39]. However, some time ago it was realized
that most of these constraints are drastically weakened by the smallness of
some Yukawa coupling constants in the standard model or its supersym-
metric extensions [40]- [42]. The general conclusion is that the initial charge
asymmetry can survive during the period when anomalous reactions are
at thermal equilibrium. Moreover, initial asymmetries in fermionic quan-
tum numbers, different from the baryon number, are usually transferred to
baryon asymmetry towards the end of the equilibrium sphaleron period.
Perhaps, the only drawback of GUT baryogenesis is that it is hardly
compatible with inBation. In inBationary cosmology there are several con-
straints on the temperature of the Universe after reheating. It should not
be larger than about 1010 GeV [43], otherwise a lot of gravitinos will be pro-
duced. However, the typical mass of leptoquarks in grand unified theories
is of the order of 1015 GeV, which is substantially larger that the reheating
temperature. Thus, there simply can not be any leptoquarks to decay and
produce the asymmetry!
There is a possible way out from this situation. It is related to the so-
called pre-heating stage of the expansion of the universe [44]. At this time
classical inBaton dynamics allows for a non-thermal production of heavy
particles because of parametric resonance [45]. Still, according to [46], the
effective leptoquark concentration nx /T 3 is typically quite small, nx /T 3 '"
10- 6 , which would require very large CP asymmetry in the leptoquark
decays, 6cp '" 1.
6. Leptogenesis
There is strong experimental evidence in favour of neutrino oscillations
[47, 48]. It neutrino oscillates, it has a mass. Theoretically, a lowest order
operator that can be added to the SM Lagrangian, has a form:
tl.L = ~ (vg¢) (¢tllj3) (5)
Jab Jv.f '
where ¢ is the Higgs doublet, Jv.f is some high energy scale, and 1I is left-
handed neutrino. This term gives Majorana neutrino masses and a lepton
number violation. A most simple way to get this effective interaction from
a renormalizable field theory is to have right-handed neutrino lIR with large
Majorana mass Jv.fR. Then (5) comes from the see-saw mechanism [49, 50].
A heavy right-handed neutrino can decay and produce lepton asymmetry
in the early universe, precisely in a way how leptoquarks produce baryon
asymmetry in GUTs. Then the lepton number is converted into baryon
asymmetry by sphalerons [51] (for a recent review see [52]). The resulting
baryon asymmetry is just a numerical factor of order one smaller than the
lepton asymmetry.
This mechanism for baryogenesis requires sufficient concentration of
right-handed neutrinos at the moment they decay. If mR '" 1010 GeV or
less, right handed neutrinos could be thermally produced at the end of in-
flationary period; the reheating temperature is sufficiently low to prevent
BARYOGENESIS 413
the overproduction of gravitino [43]. Right-handed neutrinos can also be

produced non-thermally at preheating [53].
Another interesting proposal is based on the use of neutrino oscillations
[54]. It requires relatively light sterile neutrino (the right-handed neutrino
can play this role).
7. AfHeck-Dine baryogenesis
The Affleck-Dine mechanism [55] takes advantage of supersymmetry. Su-
persymmetric theories contain scalar fields that carry lepton or baryon
numbers and the effective potential for squarks and sleptons has flat di-
rections, i.e. the energy of the static scalar field configuration at large ¢
is much smaller than ¢4. In this scenario, a combination of squarks and
sleptons, or some other fields carrying a baryon or lepton number, has a
large expectation value along some flat direction of the potential at the end
of inflation. At large VEV, the baryon number can be strongly violated by
the high-scale physics. As a result of the baryon number non-conservation,
along with the CP violation, the scalar condensate acquires a baryon num-
ber and the complex scalar field is characterized by the time-dependent
phase, ¢ = 1¢(t)1 exp(Q(t)). The subsequent evolution leads it into the
domain of conserved baryon number, because the amplitude of field ¢ de-
creases with time. Finally, squarks decay into ordinary quarks and release
baryon number stored in the scalar condensate. A study of this scenario in
different models was made in refs. [55, 56, 57] with the result that baryon
asymmetry of the universe can be explained by this mechanism.
In ref. [58] it was realized that an initially spatially-homogeneous scalar
condensate that carries baryon number may become unstable with respect
to small coordinate-dependent perturbations and develop a spatial pat-
tern that comprises domains of high and low charge density. The domains
with high charge density eventually evolve into non-topological solitons (Q-
balls), that carry baryonic number. Depending on how supersymmetry is
broken, Q-balls may be absolutely stable [58, 59] or decay into fermions at
a later time [60, 61]. In the first case Q-balls may play the role of dark mat-
ter; relic Q-balls leave a spectacular experimental signature [62] and may be
searched for in experiments like Superkamiokande, AMANDA, ANTARES
and Baikal. First limits on their flux have been already derived in the Baikal
Girlyanda experiment [63].
8. Conclusions
The baryon asymmetry of the Universe is just one number, just one (of
course, very important for our existence) characteristic of the Universe.
Why do we care so much about it? Quite interestingly, the requirement that
particle physics and Big Bang theory should be able to explain it provides
a number of interesting and strong constraints on the particle theory.
We could say that we have a qualitative understanding of the origin
of the baryon asymmetry. The most important nontrivial requirement is
that baryon number must not be conserved. Although this appears to be a
trivial statement nowadays, the baryon asymmetry of the Universe is in fact
the only observational argument in favour of B-nonconservation. A particle
physics origin of B-nonconservation is still not established. Certainly, there
were baryon-number-violating interactions operating at temperatures well
above 100 GeV - these were anomalous electroweak reactions - but whether
the baryon asymmetry came entirely from this source or was a combined
effect of the electroweak processes and grand unified and/or intermediate
scale lepton number violating interactions is unclear.
Electroweak baryogenesis is a viable hypothesis, based on a mechanism
of B-violation that is present in the standard model, and thus is very ro-
bust. Other necessary ingredients of EW baryogenesis are first-order phase
transition and CP violation, both also present in the standard model. The
electroweak theory has all the necessary ingredients for the production of
baryon asymmetry. The minimal standard model has too weak a phase
transition to produce baryons, and this points out in the direction of new
physics. This could be a "low-energy" new physics, such as supersymmetry,
or "high-energy" new physics, such as grand unification. There are several
extensions of the EW theory, which can produce the observable baryon
asymmetry, and a number of their predictions are testable in the near fu-
ture.
In any case, the explanation of the existing baryon asymmetry requires
at least a mild extension of the Minimal Standard Model. One needs strong
experimental input to solve one of the most challenging problems in cos-
mology. If baryon asymmetry has an entirely electroweak origin then new
physics should be discovered at the low-energy scale. It would show itself in
new light particles and new sources of CP-breaking, leading to CP-violating
B-decays as well as electric dipole moments of neutron and electron. Pos-
sible discovery of lepton number violation and/or proton decay would be a
strong indication of the early origin (T» 1 TeV) of the baryon asymmetry.
References
1. P.A.M. Dirac, Nobel Lecture, 1933.
2. G. Steigman. Annu. Rev. Astr. Ap., 14:339, 1976.
3. A.G. Cohen, A. De Rujula and S.L. Glashow, Astrophys. J. , 495:539, 1998
4. J.H. Christenson, J.W. Cronin, V.L. Fitch, and R. Turlay. Phys. Rev. Lett., 13:138,
1964.
5. A.A. Penzias and R.W. Wilson. Astrophys. J., 142:419, 1965.
6. G. Gamov. Phys. Rev., 70:572, 1946.
BARYOGENESIS 415
7. K.A. Olive and D.N. Schramm. Phys. Rev., D54:109, 1996.

8. A. D. Sakharov. JETP Lett., 6:24, 1967.
9. V.A. Kuzmin. Pisma ZhETF, 13:335, 1970.
10. M.E. Shaposhnikov, Contemp. Phys. ,39:177, 1998
11. V.A. Rubakov and M.E. Shaposhnikov. Phys.Usp., 39:461, 1996 (Usp. Fiz. Nauk
166:493, 1996).
12. P. Arnold, D.T. Son, L.G. Yaffe. hep-ph/9810216.
13. P. Arnold, L.G. Yaffe. Phys.Rev., D57: 1178, 1998.
14. A.G. Cohen, D.B. Kaplan, and A.E. Nelson. Annu. Rev. Nucl. Part. Sci. , 43:27,
1993.
15. M.E. Shaposhnikov, Lectures given at ICTP Summer School in High Energy Physics
and Cosmology, lheste, Italy, Jun 17 - Aug 9, 1991.
16. D.A. Kirzhnitz. JETP Lett., 15:529, 1972.
17. D.A. Kirzhnitz and A.D. Linde. Phys. Lett., 72B:471, 1972.
18. A.G. Cohen, D.B. Kaplan, and A.E. Nelson. Phys. Lett., B245:561, 1990; Phys.
Lett., B263:86, 1991; Nucl. Phys., B349:727, 1991.
19. L. McLerran, M. Shaposhnikov, N. Turok, and M. Voloshin. Phys . Lett., B256:451,
1991;
M. Dine, P. Huet, R. Singleton, and L. Susskind. Phys. Lett., 257B:351, 1991.
20. M. Joyce, T. Prokopec, and N. Turok. Phys. Rev. Lett., 75:1695, 1995; Phys. Rev. ,
D53:2930, 2958, 1996.
21. G. Farrar and M. Shaposhnikov . Phys. Rev. Lett., 70:2833, 1993; Phys. Rev.,
D50:774, 1994.
22. J.M. Cline, M. Joyce, K. Kainulainen. Phys.Lett., B417:79, 1998.
23. A. Riotto. Phys.Rev., D58:095009, 1998.
24. M. E. Shaposhnikov. JETP Lett., 44:465, 1986; Nucl. Phys., B287:757, 1987.
25. K. Kajantie, M. Laine, K. Rummukainen, and M. Shaposhnikov. Nucl. Phys.,
B458:90, 1996; Nucl. Phys., B466:189, 1996; Phys. Rev. Lett.,77:2887, 1996.
26. K. Rummukainen, M. Tsypin , K. Kajantie, M. Laine, M. Shaposhnikov. hep-
lat/9805013.
27. The LEP working group for Higgs boson searches, OPAL TN-614 (1999) (Tampere
6-49)
28. M. Carena, M. Quiros and C.E.M. Wagner. Phys. Lett., B380:81, 1996;
M. Laine. Nucl. Phys., B481:43, 1996;
J. M. Cline and K. Kainulainen. Nucl. Phys., B482:73, 1996;
M. Losada. Phys. Rev., D56:2893, 1997;
J .R. Espinosa. Nucl. Phys., B475:273, 1996;
D. Bodeker, P. John, M. Laine and M.G. Schmidt. Nucl. Phys., B497:387, 1997;
M. Laine, K. Rummukainen. Phys.Rev.Lett.80:5259, 1998.
29. M. Carena , M. Quiros, C.E.M. Wagner. Nucl.Phys., B524: 3, 1998.
30. S. Nasser and N. Turok. hep-ph/9406270.
31. M. Carena , M. Quiros, A. Riotto, I. Vilja, and C.E.M. Wagner. Nucl. Phys.,
B503:387, 1997.
32. J.M. Cline, M. Joyce, K. Kainulainen, talk at the CERN workshop CAPP-98, 1998.
33. A.Yu. Ignatiev, N.V. Krasnikov, V.A. Kuzmin, and A.N. Tavkhelidze. Phys. Lett.,
76B:436, 1978.
34. S. Weinberg. Phys. Rev. Lett., 42:850, 1979.
35. S.Yu. Khlebnikov and M.E. Shaposhnikov. Nucl. Phys., B308:885, 1988.
36. M. Fukugita and T. Yanagida. Phys. Rev., D42:1285, 1990.
37. J.A. Harvey and M.S. Turner. Phys. Rev., D42:3344, 1990.
38. A.E. Nelson and S.M. Barr. Phys. Lett., B246:141, 1990.
39. W. Fischler, G.F. Giudice, R.G. Leigh, and S. Paban. Phys. Lett., B258:45, 1991.
40. L. E. Ibanez and F. Quevedo. Phys. Lett., B283:261, 1992.
41. B.A. Campbell, S. Davidson, J. Ellis, and K. A. Olive. Phys. Lett., B297:118, 1992.
42. J. M. Cline, K. Kainulainen, and K. A. Olive. Phys. Rev., 049:6394, 1994.
43. S. Weinberg, Phys. Rev. Lett. Phys. Rev. Lett., 48:1303, 1982;
J.Ellis, A.D. Linde and D.V. Nanopoulos, Phys. Lett., 118B:59, 1982.
44. L. Kofman, A. Linde and A. A. Starobinsky, Phys. Rev. Lett., 73:3195, 1994. Phys.
Rev., D56:3258, 1997.
45. E.W. Kolb, A. Riotto and I.I. Tkachev. hep-ph/9801306.
46. S. Khlebnikov, L. Kofman, A. Linde, I. Tkachev. Phys.Rev.Lett., 81:2012, 1998.
47. Y. Fukuda et al. [Super-Kamiokande Collaboration], Phys. Lett., B436:33, 1998
48. Y. Fukuda et al. [Super-Kamiokande Collaboration], Phys. Rev. Lett.,81:1158, 1998
49. T . Yanagida, Report KEK-79-18 (1979)
50. M. Gell-Mann, P. Ramond and R. Siansky, in "Supergravity" (North-Holland, Am-
sterdam, 1979).
51. M. Fikugita and T. Yanagida, Phys. Lett., B174:45, 1986
52. W. Buchmuller and M. Pliimacher, hep-ph/990431O
53. G.F . Giudice, M. Peloso, A. Riotto and I. Tkachev, JHEP 08 (1999) 014, hep-
ph/9905242.
54. E.K. Akhmedov, V.A. Rubakov and A.Y. Smirnov, Phys. Rev. Lett., 81:1359, 1998
55. I. Affleck, M. Dine. Nucl. Phys., B249:361, 1985
56. M. Dine, L. Randall and S. Thomas. Nucl. Phys., B458:291, 1996
57. A. de Gouvea, T. Moroi and H. Murayama, Phys. Rev., D56:1281, 1997
58. A. Kusenko, M. Shaposhnikov. Phys.Lett., B418:46, 1998.
59. G. DvaIi, A. Kusenko, M. Shaposhnikov. Phys.Lett., B417:99, 1998.
60. K. Enqvist, J. McDonald. Phys.Lett., B425:309, 1998.
61. K. Enqvist, J. McDonald. hep-ph/9803380.
62. A. Kusenko, V. Kuzmin, M. Shaposhnikov, P.G. Tinyakov. Phys.Rev.Lett., 80:3185,
1998.
63. I.A. Belolaptikov et al. astro-ph/9802223.
COSMOLOGY AND DARK MATTER
PIERRE SALATI
Laboratoire Ancilevien de Physique TMorique LAPTH
B.P. 110
74941 Annecy-le- Vieux Cedex, France
and
Universite de Savoie
B.P. 1104
73011 Chambery Cedex, France
salati@lapp.in2p3.fr
Abstract. Modern cosmology is in a complete upheaval. A host of new

methods has recently flourished that allow to probe our universe and deter-
mine its geometry. The search for remote supernovae SNela supplemented
by the measurements of the cosmic microwave background anisotropies will
soon determine - within the next two decades - the cosmological parame-
ters such as OM and 0A. Observations already indicate that rv two thirds
of the universe are filled with what may be interpreted as vacuum energy
or quintessence. The remaining third would be made of non- relativistic
matter comprising mostly an important non-baryonic component. Since
Zwicky's determination of the dynamical mass of the Coma cluster in 1933,
the astronomical dark matter puzzle has challenged many astronomers and
is still awaiting for its solution. The presence of large amounts of unseen
material around galaxies and inside clusters is now confirmed by various
searches that follow independent strategies. The virial velocities as well
as the X-ray emission profiles of clusters will be discussed together with
the developing field of gravitational lensing. The evidences for a signifi-
cant amount of matter concealed around galaxies is not disputable. The
intra- cluster gas fraction may be combined to the results of the primordial
nucleosynthesis theory to yield a density of non- relativistic matter on the
order of one third - in fairly good agreement with the above mentionned
results. This cosmic concordance between various independent cosmolog-
ical tests strengthens the conclusion that OM rv 1/3 and makes a strong
case for a cosmological constant or for the presence of vacuum energy. The
417
J-J Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments. 417-510.
418 PIERRE SALATI
nature of the astronomical dark matter remains nevertheless an open ques-

tion. A small baryonic component is possible even if it does not seem to be
in the form of Jupiter-like objects or faint stars. Cold clouds of molecular
hydrogen are a viable option. The bulk of the dark matter is actually made
of a non-baryonic species. The massive and weakly-interacting particles
predicted in the framework of the putative supersymmetry are quite fash-
ionable candidates. These neutralinos may contribute a significant fraction
to the density budget of the universe. The signatures associated to their
presence inside the halo of the Milky Way will be discussed. The detection
of such evanescent signals is extremely challenging and has already led to
spectacular experimental developments in space and in the deep ocean.
1. The cosmological parameters nM and nA-

On scales smaller than'" 30 h- 1 Mpc 1 , galaxies cluster in very large struc-
tures with a quite peculiar filamentary aspect associated to the existence of
voids. The large scale structure of the universe is astonishingly reminiscent
of a brain tissue. However, on larger scales, no special structure is identified
and matter seems to be uniformly spread. The distribution of galaxies seems
also to be isotropic, i. e., it does not depend on the direction in the sky. This
observation is supplemented by the existence of a microwave background
radiation whose angular distribution is also extremely isotropic. Temper-
ature fluctuations do not exceed 10- 5 in relative value - see section 1.3.
According to general relativity, the geometry of the universe is the trans-
lation of its matter content. It should therefore reflect the homogeneous
and isotropic distribution that is observed. As a consequence, the usual 3D
space is maximally symmetric. This leads to the Robertson-Walker metric
[1,2,3,4]
which provides the geometrical framework for our description of the stan-
dard Friedmann-Lemaitre model of cosmology. Some remarks are in order
at this stage. To commence, the time interval dt that a clock at rest in-
dicates is actually equal to the proper time interval dr. Our coordinate
system is therefore in free fall - or in free expansion - and the cosmological
time t behaves exactly as if gravity did not act. Then, the space coordinates
1The Hubble expansion parameter is denoted by h when it is expressed in units of

100 km S-1 Mpc- 1 .
COSMOLOGY AND DARK MATTER 419
r, 0 and <p allow to fix the position of a star or a galaxy with respect to a
cosmic mesh that follows the expansion of the universe. The latter evolves
according to the variations in time of the scale factor a(t) . Galaxies move
away from each other not only because of their proper velocity, but essen-
tially because the cosmic network on which they sit inflates. At any time
t, the expansion rate is defined as
dIna a (2)
H(t) = dt a
The angular coordinates 0 and <p have their usual meaning. The Robertson-
Walker metric behaves as if the 3D space was flat as regards the small
displacements that are perpendicular to the line of sight. In particular, the
energy radiated isotropically by a point- like source spreads on a sphere
whose surface is given by the conventional S = 47ra 2 r2 relation. Notice
however that the radius of that sphere, i. e., the distance from the origin 0 to
any point M is equal to the product a(t) x r only in the situation where the
universe is flat and has a vanishing curvature k = 0 - see Fig. 1. In that case,
the meridian surfaces which are obtained by fixing the longitude <p at some
arbitrary value are actually 2D planes. Varying the colatitude 0 at fixed r
amounts for M to follow a circle from the north pole N (0 = 0) towards the
south pole S (0 = 7r). When the curvature k > 0, the meridian surfaces are
no longer flat. Each of those may be understood as the 2D surface of a 3D
hypersphere extending towards a non-physical fourth spatial dimension.
The origin 0 of the coordinate system becomes now the north pole of that
hypersphere. The points Nand S are just lying on a parallel of latitude
that circles around 0 on the hypersphere and comprises a bulging surface
whose area is larger than the Euclidean value of 7ra 2 r2. The universe has
a spherical geometry. It is closed as a traveller going straight ahead along
a line of sight comes back to the departure point. As a consequence, the
universe has also a finite volume.
Problem nO 1 - Level [1] : Interpret r as the radius of the above-mentionned

parallel of latitude once it is projected on the equatorial plane of the hy-
persphere (a = 1). Assume that the radial coordinate is increased by the
amount dr. Show that M is shifted by the distance
ds = dr (3)
~
The geometry of space is even more complicated in the hyperbolic case

where k < O. The surface encompassed by a parallel of latitude is now
420 PIERRE SALATI
smaller than in the Euclidean case. It cannot be embedded as a subset of a

higher-dimensional flat space. The universe behaves as if each of its points
was surrounded by a saddle. Note finally that the universe is infinite for
both the flat and hyperbolic cases.
The Robertson- Walker metric element (1) is associated to affine con-
nections whose non-vanishing elements are given by
(4)
rr kr
q;tI> = - (1 - kr2) r sin2 e,
rr 1 - kr2
(5)
r rOo -- r OrO -- r¢rtl> -- r¢tl>r -_ -1 (6)
r
and
r $4> = e e
- sin cos (7)
Problem nO 2 - Level [3] : Starting from a metric element dT2, an effective

Lagrangian may be denned as
f'
J...- = .
9J.t1l X
J.t • II
X , (8)
where i;J.t = dxJ.t / dp is the derivative of the variable xJ.t with respect to the
parameter p. The latter plays in the problem the role of an effective time.
Show that the four equations of motion associated to £, namely
f)£
d (f)£) (9)
dp f)i;J.t f)xJ.t
lead to the equation of the geodesics
(10)
Starting from the Robertson-Walker metric, readily derive the affine con-
nections (4), (5), (6) and (7).
z
N
M
r e
0Y--___------,+-_y
s
The flat case corresponds to k = O.
:0
I
The spherical geometry is associated to k > O.

Figure 1. The spherical coordinates r, () and ¢ have their usual meaning when space is
fiat - k = O. In particular, the meridian surfaces for which the longitude ¢ is constant,
are mere planes where the position of M is described by the colatitude () and the radius r.
When the universe becomes spherical with k > 0, each of those meridian planes becomes
the polar cap of an hypersphere whose north pole is now the origin 0 of the coordinate
system.
422 PIERRE SALATI
Problem nO 3 - Level [3] : The Riemann-Christoffel curvature tensor may

be expressed in terms of the affine connections r ~{3 as
R~{3v = ovr%Jl - O{3r::Jl + r~{3r::TJ - r~vr%TJ ' (11)

whereas the Ricci tensor corresponds to the contraction RJlv = R~Qv. Show
that the only non-vanishing elements of the Ricci tensor are
a .. k }
Roo = 3-
a
and Rij = gij { ~ + 2H2 + 2 a 2 . (12)
1.1. THE COSMOLOGICAL CONSTANT A.
Einstein's equations relate the geometrical content of the universe as de-

scribed by the Ricci tensor RJlv to the stress-energy tensor TJlv of the matter
at the origin of the gravitational fields
1 87rG
RJlv - :2 gJlV R - AgJlv = - 7 TJlv (13)
Note the presence of the additional term A gJlV which Einstein introduced
in his equations as a modification to the original relations. The so-called
cosmological constant A may be understood in two different ways. First, it
may be considered as another constant of mother Nature just like Newton's
constant of gravity G. If so, it should remain in the left-hand side of Eq. (13)
which deals with the geometry of space-time. We expect A to have the
same dimensions as the inverse of a length squared L"A2. That cosmological
constant should have moderate effects on the expansion of the universe.
In particular, it should come into play only in the recent past under the
penalty of playing havoc with the expansion of the universe. Depending
on its sign, a too large cosmological constant would make the universe
recollapse instantaneously or alternatively would desperately inflate space.
Setting A equal to the present value of the closure density 2 leads to
(14)
to be compared to the Planck scale
hG}1/2
Lp = { 7 = 1.6 X 10- 33 em (15)
2See below for the definition of the closure density pS.

In that approach, both Lp and LA would be the fundamental scales of

gravity. The alternative is to interpret the cosmological constant as the
stress-energy tensor of an additional fluid that fills up the universe together
with baryons and photons. The tensor A 9fll/ should therefore appear in the
right-hand side of Eq. (13) as a component of the stress-energy tensor to
which it would contribute a term
(16)
The stress-energy tensor of a gas with energy density p and pressure P is

given by
(17)
where Ufl = dxfl /dT denotes the 4- velocity. The cosmological constant may
be pictured as a fluid at rest 3 whose pressure PA and energy density PA
are given by
(18)
When A > 0, the energy density is positive whereas the pressure is astonish-
ingly negative. If the fluid is homogeneously spread, that negative pressure
gives a positive work whenever expansion occurs. Its main effect on large
scales is to pull the fluid apart and to generate inflation. The universe gains
energy as it expands under the action of PA. Whereas an ordinary gas cools
down adiabatically as it expands, the fluid under concern here has exactly
the opposite behaviour. Equally strange is its action on matter when its dis-
tribution is now inhomogeneous. When concentrated in a region of space,
the negative pressure PA acts as a source of confinement for the particles
therein. The old bag model for hadrons is actually based on the existence
inside nucleons of a non-vanishing energy density B > 0 associated to the
confining pressure PA = - B. The latter cancels the Fermi pressure of the
quarks which therefore get trapped under the action of the bag constant
B. Because of those unusual properties, such a fluid whose stress-energy
tensor is given by relation (16) is called the quintessence or fifth element as
it significantly departes from what we commonly experience. A pedagogical
illustration of the quintessence is provided by a simple neutral scalar field
r.p whose Lagrangian is
£ (19)
3Remember that the Robertson-Walker set of coordinates is in free fall. It is no

surprise therefore if the matter-energy content of the universe appears at rest in that
frame.
424 PIERRE SALATI
If that scalar field varies only in time but not in space, its energy density
obtains from the stress-energy tensor element
TOO = 1 .2
P = -<p
2
+V (20)
whereas the pressure is given by
(21)
If the potential V dominates over the time-dependent term, the scalar field
<p behaves exactly as the quintessence and its pressure P ::: - V is negative
whenever the energy density p::: V is positive.
Problem nO 4 - Level [2] : Derive the stress-energy tensor associated to the

scalar field Lagrangian of Eq. (19) where the metric 9J-Lv is not necessarily
Minkowskian and show that
(22)
Recover relations (20) and (21) for an homogeneous distribution in flat

space with V<p = O.
We are ready now to derive the equations that describe the evolution
of the universe. We will assume that the cosmic fluid is described by the
stress-energy tensor of Eq. (17) to which both matter and a cosmological
constant - or quintessence - contribute
P = PM + PA (23)
Matter is assumed to be non-relativistic with a vanishing pressure whereas

PA = -PA· As matter is diluted by the expansion, PM decreases like a- 3
while PA is constant.
Problem nO 5 - Level [2] : Derive the evolution of the scale factor a(t)
from the Ricci tensor (12) and from relation (17). Show that the time-time
component of Einstein's equations (13) leads to
87fG k
--P- (24)
3 a2 '
whereas its space-space components imply

a 47rG
- - {p + 3P} (25)
a 3
Units are such that c = 1.
The evolution of a is completely determined by the fundamental rela-

tion (24). At the present epoch, this equation may be expressed as
2 87r G 0 k 87r G 0
Ho = --P --Pc (26)
3 ao
2 3
This relation defines the density pg

as the critical value for which the
universe is fiat (k = 0). The present value p~ of the matter density may be
conveniently expressed in units of that critical density pg
n _ 0 / 0 (27)
HM - PM Pc·
The energy density PA associated to the presence of a non-vanishing cos-

mological constant leads to
(28)
The quantities nM and nA are fundamental parameters as regards the

evolution of the universe and its geometry. They will soon be determined
to a few percent accuracy by two methods which will be discussed in the
forthcoming sections: the supernovae searches (1.2) and the determination
of the CMB anisotropies (1.3) . The curvature nK may be defined as
(29)
The universe is spherical whenever nK < O. This corresponds to the portion

above the line nA = 1 - nM in the nA versus nA plane of Fig. 6. The
evolution of the scale factor a(t) may be reduced to a simple problem of ID
mechanics. It is actually equivalent to determining the position x = a( t) / ao
of a point-like particle moving on an axis as a function of the time T = Ho t.
Problem nO 6 - Level [1] : Show that relation (24) may be written as
1;2 + V(x) = E , (30)

426 PIERRE SALATI
where i; = dx/dr and E == OK . Derive the effective potential to which the

point-like particle is subject
V(x) (31)
E>O
Expansion is forever
V (x) ....... --- ~ ...-~ - --- -~ - - .... ~ .
1
x
E<O
Recollapse
.Q =0
A
No cosmological constant
Figure 2. This case corresponds to a vanishing cosmological constant with nil. = O. The
fate of the universe depends on the matter content nM = 1 - E . A negative value of E
corresponds to a trajectory that crosses V. The particle bounces backward and recoil apse
is inevitable. For E ~ 0, expansion takes place forever as the energy E always exceeds
the potential V .
Relation (30) expresses the conservation of the energy of the representative

particle. The evolution of x depends on the shape of the effective potential
V(x) and several cases are possible as OA varies. In the case of a vanishing
cosmological constant OA = 0, the potential V is Keplerian. The evolution
of the scale factor is featured in Fig. 2. The fate of the universe depends
on the sign of OM. For OM > 1 or E < 0, the universe is spherical and will
recollapse in the future. The scale factor a follows a cycloid as time t goes
on. When OM = 1 or E = 0, the universe is flat and expansion takes place
forever. Finally, for OM < 1 or E > 0, expansion is also endless while the
geometry is hyperbolic. Note that by definition OM is always positive.
V(x)
Recollapse
.......................................
Negative Cosmological Constant
Positive Pressure
Figure 3. If the cosmological constant is negative, the corresponding vacuum pressure

PI\. is positive. This inhibits the expansion and the universe always recoil apses after a
period of expansion .
When the cosmological constant is negative, the potential V increases

from - 00 for x = 0 up to + 00 when x is large. Whatever the value of
the energy E, the trajectory bounces back on the potential V and the
universe recollapses. It nevertheless may reach its present size x = 1 where
V = E - 1. In the case illustrated in Fig. 3, the values of OA and OM are
such that the potential vanishes at x = 1 while E = 1. When the values of
OM and (-OA) are too large, the potential V(x) becomes unbearably steep.
The motion of the representative particle occurs so rapidly that the age of
the universe is smaller than the lower bound of rv 8 Gyr set by galactic
white dwarves - see below the discussion of Eq. (39). Finally, for positive
OA, the potential V reaches a maximum of
3_ (") 2/3 (") 1/3

TT
vMax = __
2 2/ 3 HM HA , (32)
for XMax = (OM/2 OA)1/3. If the effective energy E exceeds VMax, our rep-
resentative particle makes it to the top of the hill and rolls down on the
428 PIERRE SALATI
other side. After a period of deceleration corresponding to x < XMax, the

expansion of the universe reaccelerates. If on the other hand E < VMax, the
universe recollapses eventually. This corresponds to the region below the
corresponding border line of Fig. 6. The expansion is reaccelerated today
Positive Cosmological Constant
Negative Pressure
v(.x) 1
x
Deceleration Acceleration
Acce .ration
Figure 4. If the cosmological constant is positive, the associated negative pressure PA

drives the expansion. After a period of deceleration during which the maximum M of
the effective potential is reached, the universe rolls down on the other side of the hill. In
that scenario, we are today, at x = I, in a stage of acceleration for which a > O.
if the condition XMax < 1 is met so that the universe is in the roll- down
phase today at x = 1. Present reacceleration corresponds to the region for
which OA > OM /2. Note finally the presence of a grey shaded region in
the upper-left corner of Fig. 6 where the No Big Bang sign shows up. In
that part of the plot, two irreconcilable conditions cannot be satisfied at
the same time. On the one hand, recollapse occurs as the energy E does
not exceed the top of the potential V. On the other hand, because that
region is above the OA = OM /2 limit, we should be today in a stage of
reacceleration and the corresponding state x = 1 should be on the other
side of the hill, a region which the representative point cannot reach since
it bounces back before. The shaded region corresponds to a configuration
where our present universe cannot exist. It is therefore forbidden.
1.2. THE SUPERNOVAE SEARCHES.
An important test of cosmology is the determination of the relationship

between the distance of luminosity dL and the redshift z. In a flat and
static space, a source with absolute luminosity L would shine with the
apparent luminosity l at distance r so that
(33)
The same relation is used in order to define dL in cosmology where the space
inflates and the geometry is not necessarily Euclidean. Let us consider a
distant source, say a quasar or a galaxy, with absolute luminosity L. That
source is at rest with respect to the expanding mesh of the Robertson~
Walker coordinates. It is located at the co~distance rl from the Earth. The
quantity of light that it emitted at the remote time tl during the time
interval otl comprises an energy oWe = LOtI. That energy is radiated
isotropically and spreads today as it reaches the Earth on a sphere whose
surface is equal to 47r rt a5. The apparent brightness l is the amount of
energy received per unit surface and per unit time. The energy oWe emitted
in the past during ot l is now received during the time interval ota such that
oto = (1 + z) = ao . (34)
Otl al
As a result of the expansion of space by a factor of aO/al, the electromag-

netic waves are stretched while they propagate at the speed of light. This
implies that all wavelengths are red~shifted and the amount oWr of energy
that is eventually collected is also reduced by the same factor of (1 + z).
Because the apparent brightness may be expressed as
l _ { oWe } oh
- 47r a5 rr 1
Oto (1 + z) ,
(35)
we infer a distance of luminosity equal to
dL = rl ao (1 + z) , (36)
for a source at co~distance rl. We would like to establish now the relation
between the distance of luminosity dL of the source and its redshift z.
430 PIERRE SALATI
Problem nO 7 - Level [2]: The relation between the reduced scale factor
x = a(t)/ao and the amount z by which the light emitted at time t is red-
shifted is presented in Eq. (34). Starting from relations (30) and (31), show
that the variation dz of the redshift may be expressed as
H dt = - dz (37)
o (1 + z) JF(z) ,
where the function F(z) is defined as
F(z) = (1 + z)2 (1 + z OM) - z (2 + z) OA . (38)
The previous relations allow to determine the age of the universe as an

integral on the redshift
to = t to
0 dt -
-
H- I
0
t+oo
0 (1 + z)
dz
JF(z)
(39)
If OM increases or if alternatively OA decreases by becoming more and more

negative, the subsequent increase of the function F(z) leads to small values
of the age of the universe. The lower-right corner of Fig. 6 becomes excluded
- see the grey shaded region - on the basis that the corresponding age is
smaller than the typical value of", 8 Gyr already mentionned. A photon
emitted at time tl is received today at to . The co- distance of the source is
ri. The propagation of light corresponds to a vanishing interval of proper
time dT = 0 so that
(40)
Problem nO 8 - Level [1] : Show that the right-hand side term of Eq. (40)
may be expressed as
~ arcsin ( v'k rl ) k > 0 (spherical)
h
rl k = 0 (fiat)
sinh- I (Fkrl) k < 0 (hyperbolic)
Show now that the left- hand side term of Eq. (40) may be transformed into
{to dt 1 (z dz'
(41)
it l a(t) = aoHo 10 JF(z')
Demonstrate finally that when the space is not flat, i.e., for k of. 0
(42)
We are now ready to derive the relation between dL and z. In fiat space,
we readily infer that
(43)
In the case of a spherical (k > 0) or hyperbolic (k < 0) universe, the

calculation is slightly less straightforward. One may first show that
d = ~(l+z) ~Iklr
L Ho JlTIKT V11\;1 I , (44)
in order to get eventually the standard relation
d _ ~ ( 1 + z) {~{Z dz' }
L - Ho JlTIKT S VIHKI 10 JF(ii) , (45)
where S(x) = sinx for a spherical geometry (k > 0) and S(x) = sinhx
when the universe is hyperbolic (k < 0). The cosmological parameters
come into play both through OK and the function F(z). If we extend the
definition of S(x) so as to be equal to x in the fiat case (k = 0), Eq. (43)
is directly recovered from the general relation (45) as the curvature term
JlTIKT cancels out. We may apply our calculations to the case where the
cosmological constant vanishes.
Problem nO 9 - Level [3] : Show that in the case where OA = 0, the

distance of luminosity may be expressed in terms of the redshift z as
(46)
432 PIERRE SALATI
where the deceleration parameter qo = DM /2.
Supernova Cosmology Project

Perlmutter et al. (1998)
~ ..
.. - -.~. ~~~
(0.0)
" (1.0)
(;.iJ (2.0)
o
II
<:
Supernova
Cosmology
Project
1.5
c
.~ 4t
6 ~~-'-~----'
... .
._ HH. >!.~~~>.:.:; HH ~H sHHHHHl
<J
~ ~r:
,~O~'H
. . __. . . . .... . . . . . ... ..._
.._
. _.__
. _..._
.. ._._
_. _
. _.. _
. ._
._
] -2 ..... 0. . ..... . .... .. ·· · · ·0··· ........ . . ....... .... . . . . ..... ......... ........ . .. J
~ :ci .HH.~ __~ ~ _
.~
0.0 0.2 0.4 0.6 0.8 1.0

redshift z
www-supernova.lBl.gov
Figure 5. Measurements of the apparent magnitude versus redshift relationship for two
samples of supernovae SNela. The yellow group corresponds to local objects while the
red data points are remote sources located at cosmological distances [5] .
Measuring the distance of luminosity to red-shift relationship allows to

determine a specific combination of the cosmological parameters DM and
DA·
Problem nO 10 - Level [1] : Relation (45) may be expanded as a series of

the redshift z in our neighborhood. By developing the function F(z) up to
first order in z, show that
1
yfF(Z) ~ 1 {nM}
+ nA2 - 1 z + ...
- (47)
When the argument x is small, notice that sin x ~ x ~ sinh x. Expand the
distance of luminosity dL up to second order in z to get
(48)
where the parameter qo is now equal to the combination nM /2
Because the measurements extend beyond our immediate vicinity - values

of the redshift are obtained up to z rv 0.8 - the actual combination to
which the supernovae measurements are sensitive is rather the difference
(n M - nA). In the upper panel of Fig. 5, the degeneracy is obvious for
(OM, nA) = (0.5,0.5) and (0, 0) as well as between the (1.5, - 0.5) and
(2,0) cases.
Measuring the distance of luminosity requires the knowledge of both the
apparent brightness I and the absolute luminosity L. The former quantity
is determined by direct observation just as the redshift is obtained from
the spectrum. The knowledge of the absolute luminosity is based on the
putative existence of standard candels whose characteristics are expected
to be independent of space and time. Supernovae SNela are believed to be
white dwarves that accumulate the gas of a companion star by accretion.
When the mass exceeds the Chandrasekhar limit of rv 1.4 M G , the elec-
trons at the center are so compressed that they become ultra-relativistic.
The adiabatic index fl of the inner material no longer exceeds the critical
value of 4/3. The white dwarf becomes unstable against hydrostatic equi-
librium and its core collapses. The density and temperature increase up to
the point where the conditions for the ignition of carbon and oxygen are
met. In the degenerate conditions of the core, there is no thermal regula-
tion to inhibit the temperature increase as the reaction starts. The latter
proceeds through the propagation of a burning front throughout the star
at subsonic velocities. Numerical simulations with such a carbon deflagra-
tion are more consistent with observations than models with a supersonic
detonation which would exclusively turn out the stellar gas into iron , in
disagreement with spectroscopic restrictions. Because the mass involved in
434 PIERRE SALATI
that deflagration cannot exceed the Chandrasekhar threshold, supernovae

of the SNela type are believed to be standard candels with reproducible
features.
Supernova Cosmology Project

Perlmutter et al. (1998)
3
No Big Bang
90% 42 Supernovae
2
1
<
a / /
,-
I I \\1..
t'\~
"
......................)y\\.
0
rccn\bp\~. c\llllU"
c/,
rial 0&
.\ - () ~Q'
~
-1 l'ni\ 'p,t! 90. ~I'
~q
0 1 2 3
QM
Ap.J.
astro-ph/9812133
Figure 6. Implications of the supernovae searches on the values of nMand nA [5J .
Observation shows that the peak luminosity reaches 1010 L 0 . The super-
nova becomes as bright as the host galaxy. It is visible at cosmological
distances. A supernova located at z = 0.4 is detectable in 10 minutes on a
2.5 m telescope. The event lasts for a month on average. Detection of such
explosions is therefore possible by continuously monitoring distant galaxies
and by comparing directly the images collected from day to day. Since the
energy output is a priori constant, the shortest explosions were expected
to be also the brightest. The opposite correlation has actually been found
and is still awaiting for the post diction of the theoreticians. The luminosity
seems nevertheless to be constant at fixed values of the duration, making
the objects reliable candels.
Observations lead directly to the apparent magnitude m and to the
redshift z of the sources . An important feature of the dL -z relation is the
possibility to disentangle the Hubble constant Ho from the other cosmolog-
ical parameters. Relation (45) may actually be written as
(49)
where the function 9 depends only on the redshift and on the parameters
OM and OA at stake. In our local neighborhood , it simplifies into
9 {z, OM, OA} -::::: z + {OA - °2M + I} z; . (50)
The apparent magnitude m depends on the luminosity l through

m = - 2.5 19lO l + C , (51)
where C is a constant. The absolute magnitude M corresponds to the
brightness of the same source should it be now shifted in space at the
canonical distance of 10 pc. The relation between the apparent and abso-
lute magnitudes is
m - M = 5 IglO dL - 5 , (52)
where the distance of luminosity dL is expressed in parsecs [pc].
Problem nO 11 - Level [1] Show that the apparent magnitude may be

expressed as the sum
m = 5 19lO {g} + M + 5 19lO {;O (10 GpC)-l} + 45 . (53)
The first term in the right hand side of Eq. (53) depends only on the redshift
and on the cosmological parameters OM and OA. The other contributions
amount to the constant quantity
M = M + 5 19lO {;o (10 GPc)-l} + 45 , (54)

436 PIERRE SALATI
which depends both on the absolute brightness M of the candels and on the
Hubble constant Ro. Investigation of local supernovae allows for the cali-
bration of their absolute luminosities M - see the yellow group of nearby
SNeIa in the measurements of Fig. 5. Their distance may actually be mea-
sured by using other estimators. Furthermore, because in our immediate
vicinity the m - z relation simplifies into
m = 5 IglO Z + M , (55)
the quantity M may be eventually determined . Analysis of a local sample

of sources leads therefore to the measurement of the absolute magnitude M
and of the Hubble constant Ro. Note that our difficulty in estimating M
is reflected by the still large uncertainty in the determination of R o. Those
difficulties do not affect fortunately the potential discovery of a departure
from the pure IglO z behaviour of the apparent brightness m as observa-
tions will probe deeper regions of space. At large distances, relation (55) is
replaced by
m = M + 5 IglO z + 5 IglO {I + (DA - D; + 1) ~ + . . . } . (56)
The second group comprising the red data points of Fig. 5 is based on the
analysis of a sample of supernovae SNela located at cosmological distance, a
region with z rv 0.4 - 0.8 where the parameters DM and DA may actually be
measured. As already discussed, the combination DA - DM is the quantity
which is constrained by the observations. That is why the allowed lightly
shaded region of Fig. 6 has this so peculiar oval shape structure with slope rv
1. The High- z supernova search team obtains DM (A = 0) = - 0.35 ± 0.18
and DM (DK = 0) = 0.24 ± 0.1 whereas the Supernova Cosmology Project
gets [5J 1.3 DM - DA ~ - 0.4 ± 0.2 and
+0.08 +0.05
. 9 (stat) _ 0.04 (syst) .
DM (DK = 0) = 0.28 _ 0 0 (57)
A flat universe filled only with matter - DM = 1 and DA = 0 - is strongly

excluded. Quite exciting is the evidence at the 30" level for a positive value
of DA • If flat, the universe would contain rv 30% of matter whereas the
cosmological constant or quintessence would amount up to 70% of the closer
density. The dark or vacuum energy would be therefore the dominant source
of gravity today.
1.3. THE COSMIC MICROWAVE BACKGROUND.
The serendipitous discovery of a cosmological microwave background is one

of the cornerstones of the hot big bang model together with the recession of
galaxies and the theory of primordial nucleosynthesis. In 1941, McKellar [6]

found that cyanogen radicals which he detected in absorption on the line of
sight towards the star ( Ophiuchi had their rotational levels excited as if the
molecules were in thermal equilibrium with a thermostat with temperature
'" 2.3 K. In 1948, Alpher and Gamow [7] were working out the scenario
according to which", 24 % of the nucleons end up into 4He as a result of
the nuclear fusions that took place in the early universe, at an age of a few
minutes. They showed that the onset of primordial nucleosynthesis occurs
as soon as deuterium stops to be photo dissociated, i.e., for a temperature of
order 109 K. They reasoned that the number density of nucleons should be
on the order of 10 18 cm -3 at that time, so that a fraction a n v '" 1/3 only of
the nucleons could interact and be transformed into heavier elements. They
were led to the conclusion that the baryon- to- photon ratio is TJ '" 10- 10
and that a residual radiation should exist today with temperature'" 5 K.
Their conclusion did not draw too much attention and remained forgotten
for two decades. In 1965, the possibility to detect the residual radiation of
the putative primordial fireball was investigated by Dicke, Peebles, Roll and
Wilkinson [8J. Roll and Wilkinson were taking seriously the possibility to
observe a black-body radiation with temperature'" 10 K. Before they could
complete such a measurement, they learned that two researchers working
for the Bell Laboratories - Arno Penzias and Robert Wilson - had detected
a weak background signal at a radio wavelength of). = 7.35 cm. Penzias and
Wilson were actually trying to calibrate the large horn antenna at Holmdel,
New Jersey, for the Echo satellite project. In spite oftheir efforts, they could
not get rid of a residual antenna noise [9]. The latter was interpreted by
their colleagues as the radiation flux corresponding - at the frequency of
the antenna of 4080 MHz - to a black-body emission with temperature
3.5 ± 1 K. Before claiming that the cosmic microwave background (CMB)
was a pure black- body, its spectrum needed to be measured on a range of
frequencies which unfortunately is partially absorbed by the atmosphere. In
1989, the cosmic background explorer (COBE) satellite was launched and
its FIRAS instrument [10J confirmed that the CMB spectrum is Planckian
to a high accuracy. This fact can only be interpreted in the framework of a
hot big bang. Today's background is understood as the fossilized radiation
of the primeval photon gas that was once in thermal equilibrium with the
rest of the cosmic plasma. More recently, analysis of the measurements by
the DMR instrument [11] on board the COBE satellite have shown clear
evidence for anisotropies in the CMB at the 10- 5 level. This observation
was impatiently expected. It indicates that at an age of'" 3 x 105 yr, the
seeds of the future galaxies are already present and that the large scale
structures grow from the gravitational collapse of initial inhomogeneities in
the matter distribution.
438 PIERRE SALATI
At very early times, matter is split up into its constituent particles. It

exists as a relativistic radiation. Around a temperature T '" 1 MeV, the
reactions between the various species occur at a rate that exceeds by many
orders of magnitude the expansion rate of the universe - see section 3.2 for
a more detailed discussion. In particular, the fact that the set of reactions
(58)
is in equilibrium implies the equality of the corresponding chemical poten-

tials
(59)
We readily infer that the chemical potential of the radiation vanishes so
that the corresponding spectrum is actually Planckian since it follows a
Bose-Einstein distribution with fJ."( = O. Below a temperature of '" 5 x 109
K, positrons and electrons annihilate. The relaxation of fJ."( towards zero and
the corresponding restoration of a Planck distribution should the spectrum
be distorted proceed through the slower bremsstrahlung
e- + p --+ e + p + 'Y , (60)
and double Compton mechanisms
(61)
Both reactions allow for a change in the number of photons. If that essential
condition is not met , the Planck shape of the radiation spectrum is no longer
protected against the potential distortions which would arise should energy
be released in the cosmic plasma. For a low baryon-to- photon ratio 'f/, the
dominant mechanism is the double Compton process whose typical time
scale is given by
t DC c:= 8a
7r (mec2)2
kTe t,,(e, (62)
where t"(e = (n e caT)-1 is the collision time scale of a photon on the free
electrons with number density n e. The Thomson-Compton cross section
aT may be expressed as a function of the classical radius of the electron
ro = e2 /4'7rm e c2 = 2.818 X 10- 13 cm
aT = 87r r~ c:= 6.67 x 10- 25 cm2 (63)

3
• As long as the age of the universe t exceeds the double Compton time
scale t DC, any distortion in the CMB spectrum is immediately erased. An
injection of heat leads to the immediate increase of the average energy
E-y per photon. Energy is shared among photons through their collisions
with the free electrons of the plasma. The much slower double Compton
process allows to increase the number density n-y so that E-y decreases. The
chemical potential relaxes accordingly towards zero and a Planck spectrum
is re-established .
• For a redshift ZPl ...... 1 - 2 X 107 , the double Compton process freezes out.
Its typical rate t DC -1 becomes smaller than the expansion rate C 1 . This
takes place rv 105 seconds after the big bang. If an amount Jp-y of energy
is injected afterwards, the chemical potential varies according to
dp = 1.4 bp-y (64)

P-y
The spectrum of the radiation follows subsequently a Bose- Einstein distri-
bution
1
fBE(x) = ( ) , (65)
e X +P -1
where the parameter p - defined as the ratio (- p-y / kTe) - may departe
from O. The photon frequency v is expressed in terms of x = hv / kTe where
Te denotes the electron temperature. The redistribution of energy among
photons is described by the Kompaneets equation
t C
aatf IK = 1
x2 ox
a {x 4 ( ox
af + f + f 2) } , (66)
The thermalization of the radiation is due to the inverse Compton scattering

of low-frequency photons by energetic electrons and is associated to the
typical time scale
t C~ (~~2) t -ye . (67)
As long as the age of the universe t exceeds t c, the right- hand side term of
Eq. (66) relaxes very rapidly towards O. As already mentionned , a release
of heat results into the Bose-Einstein distribution (65) and the chemical
potential p departes from zero. The radiation spectrum is severely distorted
with respect to a Planck function .
• Below a critical redshift ZBE rv 2 X 105 corresponding to an age of rv 109
seconds, the thermalization of the photon bath through its interactions
with electrons decouples from equilibrium. The inverse Comptonization
time scale t c becomes larger than the expansion time t. This has two
important consequences. Any direct release of photons - for instance by an
hypothetical heavy particle that would radiatively decay - distorts imme-
diately the radiation spectrum. Alternatively, if energy is now injected in
the electron-proton gas, the electron temperature Te increases with respect
440 PIERRE SALATI
to the radiation temperature T.y• As photons collide with the hot electrons
and gain energy, high frequencies are populated at the expense of low en-
ergies. The resulting radiation spectrum exhibits an excess in the Wien
region related to a deficiency in the Rayleigh-Jeans regime. The distorted
Rayleigh-Jeans temperature decreases by a factor of e- 2y
(68)
where the evolution of the parameter y follows the equation
y = J {Te - T,} dt .
Te tc
(69)
As already mentionned, photons no longer thermalize with each other through

their interactions with electrons. Note however that the baryon population
remains strongly coupled to the radiation. The electron temperature Te
relaxes towards the photon temperature T, with a typical time scale
t e, -_ -3 {me
--
c} -
x;l- , (70)
4 O"T P,
which remains orders of magnitude smaller than t ,e'The energy density of

the photon bath is denoted by P" As long as the primordial plasma remains
ionized, the fraction Xe of free electrons is unity and baryons are subject to
follow the thermal evolution of the radiation thermostat .
• At some point, electrons and protons combine into neutral hydrogen
atoms. This so-called recombination implies the decoupling of the baryons
from the radiation background. As the ionized fraction Xe sharply drops, the
e,
time scale t jumps above t. This has important consequences on the inhi-
bition of structure formation before recombination as baryons are dragged
by the radiation. As soon as recombination has taken place, baryons behave
as an independent fluid. This occurs,....., 3 X 105 yr after the big bang.
Problem nO 12 - Level [1] : Let us consider a non-relativistic perfect gas of

particles with mass m. Show that its density n is related to its temperature
T and chemical potential fJ. through
mkT}3/2
n = 98 { 211" 1i 2 exp {(fJ. - mc2 ) / k T} (71)
What does 98 represent ?

Problem nO 13 - Level [1] : Show that the density of a photon bath with
temperature T and vanishing chemical potential is given by
n, =
2
7[2 ((3)
{kheT}3 ' (72)
where ((3) = 1.20205. Compute the number density n~ of photons today.
COBE has set stringent constraints [12J on the potential distortions of

the CMB spectrum. To a high-precision, the radiation has a black- body
emission with temperature
To = 2.728 ± 0.004 K (73)
The current limit on a y -distortion is
y :::; 1.5 X 10- 5 @ 95% , (74)

while the bound on the chemical potential parameter f-L is
f-L :::; 9 X 10- 5 @ 95% . (75)
Problem nO 14 - Level [2] : In this problem, we would like to estimate at

what time recombination takes place. Electrons and protons combine into
hydrogen atoms through the reaction
(76)
Assuming that statistical equilibrium holds, show that the free electron
fraction Xe satisfies the relation
Kx
_ x; _
- 1 _ xe -
2- 5/ 2 V1f {me c2 }3/2
T/ ((3) kT e
_ B /T
, (77)
where T/ = nB/n, '" 10- 10 . The Rydberg energy is B. Transform Eq. (77)
into
Kx= '119 X 1018 T4 -3/2 ,/10 -1
'Y) e-15.76/T4 , (78)
where T4 is the temperature in units of 104 K. Compute the transition
temperature for which Xe = 0,5 assuming a baryon-to-photon ratio equal
442 PIERRE SALATI
to "710 = 1. Show that recombination occurs at a redshift of rv 1300. What

is the corresponding age of the universe ?
Before recombination, a photon undergoes '" a hundred collisions on

free electrons during a typical expansion time. The primordial plasma is
therefore opaque. It behaves like a cloud in which light diffuses. As recom-
bination takes place, the ionized fraction Xe drops and free electrons be-
come scarce. The Thomson-Compton interaction time scale t ')'e increases
significantly. The plasma becomes transparent as it recombines. Today's
microwave background has been emitted at that time and has propagated
undisturbed since then. It is a snapshot of our universe'" 3 x 105 yr after
the big bang, in a very early stage of its evolution. Anisotropies in the CMB
correspond to the presence of inhomogeneities in the plasma distribution at
the time of recombination. Those perturbations have grown later on to form
the galaxies and their clusters of the present epoch. As seen from the Earth,
the cosmic background radiation originates from the surface on which its
photons have scattered for the last time. Several mechanisms are responsi-
ble for the temperature fluctuations seen on that so-called last scattering
surface. Light from overdense regions undergoes a gravitational red-shifted
- the Sachs- Wolfe effect - as it escapes from the potential wells. The cor-
responding temperature variation is related to the potential perturbation
 e < 0 through
6T = e (79)
T
An overdense region is also associated to a hot spot as the plasma is more
compressed therein than on average. The corresponding temperature in-
crease leads to a blue-shift and tends to cancel the Sachs-Wolfe contribu-
tion.
Problem nO 15 - Level [1] : Let us consider a fluid where the pressure

P and the energy density p are related through P = a p. The universe
expands adiabatically. As the scale factor a increases, show that
p ex: a -3 (1 + a) . (80)
Neglecting the curvature of the universe, relate the age t to the scale factor
a
t ex: a 3 (1+a)j2 . (81)
Time passes by more slowly in an overdense domain. This may be inter-

preted as if the evolution of the universe was slightly delayed with respect
to what happens out of the perturbation. The corresponding time shift ot
is of gravitational origin
Jt _ (82)
t - e
It is related through Eq. (81) to the perturbation oa of the scale factor

Jt 3 oa
t = 2" (1 + a) 7; , (83)
as if space had expanded to a lesser degree than elsewhere. As the wave-

length of photons is proportional to a, we infer the frequency perturbation
oT ov oa (84)
= =
T v a
An overdense region is therefore hotter by an amount
oT = _ ~ 1 (85)
T 3 (1 + a) e
Problem nO 16 - Level [1] : Add both contributions (79) and (85) together
to infer the global temperature fluctuation
oT _ (3a + 1) 
(86)
T - (3a + 3) e'
The Sachs- Wolfe effect is the dominant source of temperature variation.

The cold spots on the sky correspond to overdense regions from which
the light is reddened by gravitation. Compute oT IT for a pure radiation
(a = 1/3) and for a non- relativistic gas (a = 0).
A Doppler shift may also arise as the emitting zone moves on the line of
sight. Finally, photons may cross an expanding or a collapsing lump of
matter. As they make their way through it to the Earth, the potential well
of the lump may have time to drastically change. This also results in a
change in the frequency of the light.
The anisotropies of the cosmic background radiation are due to the
presence of inhomogeneities in the distribution of the primeval plasma.
Those density fluctuations are the seeds of the future galaxies. They behave
actually as mere sound waves that collapse under their own gravity while
444 PIERRE SALATI
the expansion of the universe tends to stretch them. Before recombination,

baryons and photons behave as a single fluid. The radiation provides by far
the dominant contribution to the pressure, the entropy and even the energy
density of the plasma. Because of the Thomson-Compton diffusion, baryons
are glued to the radiation that behaves as a thermostat and prevents the
baryons from collapsing alone.
II
~ 20
(Compiled by M Tegmark, Nov .

o
10 100 1000 10 4
Multipole
Figure 7. A compilation by M. Tegmark of the various CMB anisotropy measurements

is presented. The temperature variations across the sky have been expanded in spherical
harmonics. The temperature difference between two directions is averaged over the sky
and is featured by the coefficients C, of the ordinate. The angular separation () between
those two directions is described by the multipole number l = 200 0 / () plotted on the
abscissa. The observations point towards the existence of the first acoustic peak. The
uncertainties are still fairly large.
Problem nO 17 - Level [2] : Show that the pressure P of the photon-baryon

gas may be expressed as
(87)
where the energy density of the radiation is
(88)
Compute the ratio PB / P,,! and conclude. Show that the entropy per baryon
may be written as
2 7r 4 k 10 1
aB = 45 ((3) ry '" 3.6 X 10 'T/I0 - k (89)
At high temperature, the dominant contribution to the energy density is

provided by the photons. As the universe expands, the radiation is red-
shifted and baryons take over. Show that this happens when
2
T = 30 1"(3) mBc (90)
7r4 .., 'T/ k '
for a temperature of'" 400 'T/1O K. As primordial nucleosynthesis favours a

value of the baryon- to-photon ratio 'T/I0 comprised between 1 and 6, could
baryons dominate the energy density before recombination ?
We would like to understand how the density perturbations of the plasma

develop in time. We will present here a heuristic approach of the problem.
A proper treatment implies an expansion of the Einstein's equations (13) in
terms of the perturbations Jg/-LV to the metric and JT/-Lv to the stress-energy
tensor which describe the fluctuations. We will considerably simplify the
problem by assuming that the primeval plasma is matter-dominated and
behaves as a non-relativistic gas whose pressure is negligible with respect to
its energy density. As shown previously, this assumption is not correct but
we will nevertheless follow it for simplicity. The properties of the primeval
fluid with energy density p, pressure P, velocity v and gravitational field 9
follow the equation of continuity
-8p + \7.
-
(pv) o, (91 )
at
446 PIERRE SALATI
the Euler equation

Dv
Dt
= ov
ot + (v . \7 ) v = 1 -
--p\lP + 9 (92)
and the gravitational field equation

\7·9 = -41fGp (93)
The gravitational field 9 is irrotational so that
\71\9=0. (94)
Following Sir James Jeans, the plasma is perturbed around a homogeneous
state that evolves in time. The energy density may be written as
p(M, t) = po(t) + Pl(M, t) . (95)

The same expansions hold for the pressure P = Po + PI, the velocity v =
Vo + VI and the gravitational field 9 = 90 + 91 . In the unperturbed state,
the energy density decreases as a- 3 whereas the velocity takes only into
account the expansion of space
a _
- r . (96)
a
In our Newtonian approach, the gravitational field is related to the mass
contained in a sphere of radius r centered on the origin of the coordinate
system
- = - --Po
go 41f G r- . (97)
3
Note that since \7 . r = 3 and \7 1\ r = 0, relations (93) and (94) are readily
satisfied.
Problem nO 18 - Level [1]: Show that the equation of continuity (91)

and the Euler equation (92) are satisfied in the unperturbed state. Derive
relation (25) - remember that Po < < Po according to our assumptions.
Problem nO 19 - Level [2] : Show that the density fluctuations follow the
linearized equations
aih a_ a (_ n) _ 1 np _ (99)
at + -;;, VI + -;;, r· v VI = - Po v 1 + 91 ,
'\7 . 91 = - 4 7r G PI , (100)
'\7 A 91 = 0 (101)
in the limit where they are small as compared to the unperturbed state.
The pressure fluctuations are related to the density perturbations through

PI = C 52 Pl. We then concentrate on the specific Fourier mode
PI (r, t) = PI (t) exp { i k. f} . (102)
The fluctuations may be seen as the superposition of plane waves with co-
moving wave vector k. In the linear regime, these plasma waves are decou-
pled from each other. Each mode behaves as if it was alone. The perturbed
quantites PI, PI, VI and 91 vary according to Eq. (102). We remark that
the operator '\7 may be formally replaced by the vector i kla. The time
derivative aI at may also be understood as the operator
a
- -7
d
-
.a-_
-2-k'r (103)
at dt a2
Problem nO 20 - Level [3] Using the above-mentionned substitutions,

derive the equations
a k . VI-
PI + 3 - PI + 2. -Po 0 , (104)
a a
and
a _
VI + - VI = . C5 2 PI
-2-- -
k + 91- (105)
a Po a
Show that the gravitational perturbation 91 is aligned on the wave vector
k
(106)
The density contrast 0 = PI! Po is not affected by the expansion of the

universe. Its increase in time accounts only for the collapse of matter and
the formation of structures. We are interested in the sound waves that
develop in the cosmic plasma. Those are compression modes for which the
448 PIERRE SALATI
motion takes place in the same direction as the propagation of the acoustic
wave. The velocity ih is therefore aligned on the wave vector k. Show that
k · ih = ia8 . (107)
Derive the equation of evolution
(108)
In a matter-dominated universe, structures develop according to the dif-

ferential equation (108) .
• In the absence of gravity and of expansion, that relation describes the
propagation of sound waves at a speed Cs. An increase in the density results
into a growth of the local pressure so that the gas bounces the fluctuation
forward.
• Let us consider now gravity while keeping the universe static. We recover
the previous behaviour as long as the wave vector k is large. However, for
big enough structures k is so small that the coefficient in front of the 8 term
becomes negative. The fluctuation no longer oscillates. It grows exponen-
tially. Large structures are dominated by their binding gravitational energy
at the expense of a negligible thermal contribution. A small perturbation in
their inner pressure induces their collapse as they are too heavy to bounce
back. The typical scale AJ above which collapse occurs is the product of
the sound speed Cs by the characteristic gravitational time scale 1/ J GPo
of the medium
AJ = Cs J
G:o . (109)
This scale is associated to the existence of the so-called Jeans mass M J ,.....,
PoA3· Only structures whose masses exceed the critical value of MJ can
form.
• Before recombination, photons provide the bulk of the energy density
and of the pressure of the plasma. As a consequence, the sound speed is
Cs = c/V3 and the Jeans length is
(110)
The size c t of the causal horizon is defined as the distance over which
photons have travelled since the bang and over which information may
have been transmitted. The Jeans length is actually of the order of the
horizon size. We infer that a perturbation which is larger than the horizon
collapses. Just before recombination, the universe is matter dominated and

its scale factor a evolves as t 2 / 3 .
Problem nO 21 - Level [1] : Compute today's value of the critical density
(111)
Show that the universe becomes matter dominated at a redshift of
(112)
The bulk of the non-relativistic matter amounts to DM 1/3 - see sec-

f'V
tion 2.2. Compute the redshift corresponding to the matter-radiation equiv-

alence for h = 0.65. Is the universe matter dominated before recombination?
Density fluctuations that are outside the horizon grow according to a sim-
plified differential equation where the moderating effect of expansion is now
added
.. a .
o + 2 - 0 - 47r Gpo 0 = 0 .
a
(113)
The general solution is a superposition of the growing and decaying modes
(114)
Structures larger than the horizon collapse with a density contrast 0 that
increases like t 2 / 3 . Their scale A is stretched by the expansion and evolves
as a ex: t 2 / 3 . The horizon scale
f'V c t grows faster. As time goes by, large
scale fluctuations that were initially collapsing enter eventually inside the
horizon where they stop developing. They oscillate as mere sound waves and
are slowly erased by expansion as a result of the a/a term in the equation
of evolution .
• At the end of recombination when T f'V 3,500 K, baryons decouple from
the radiation. The sound speed significantly drops
1O kT}I/2
Cs = { - - f'V 3.3 X 10- 5 c (115)
3 mB
The associated Jeans length is 80 light years. We infer a Jeans mass of
f'V
3 x 10 1]10 M 8 . With 1]10 in the range 2 - 6, the first structures to form

f'V
4
seem to be the globular clusters with a mass 0.6 - 2 x 105 M 8 .
f'V
The cosmic background radiation is a snapshot of the universe when pho-

tons decouple from matter. The anisotropies reflect the presence of density
450 PIERRE SALATI
perturbations that grow or oscillate depending on their size. The critical

angle fJ H '" 10 - 20 sustained today by the horizon at recombination plays a
key role. For large angles - above fJ H - we are just observing the slow growth
of structures that exceed the horizon size. The variation of the temperature
fluctuations with angular scale reflects the dependence of the density con-
trast <5 on the comoving scale k. Fluctuations smaller than fJH correspond
to modes that are already inside the horizon and that oscillate. The epoch
at which a specific mode starts to behave as a damped acoustic wave is
directly related to its size. Varying the angular scale allows therefore to
explore the entire palette of the oscillation patterns. As a mode enters the
horizon and vibrates, its temperature oscillates and goes through a series
of maxima and minima which are gradually erased by the expansion. This
trend is imprinted on the CMB where each angular scale corresponds to a
specific stage of that evolution as it gets frozen by the sudden decoupling
between baryons and photons.
0.6
0 .6
0 .4
0.2
BOOMERANG/NA
+COBE
0.0 --L-..-.
0.0 0 .2 0.4 0 .6 O.B 1.0

Multipole. I
Figure 8. The recent measurements of the eMB fluctuations by the Boomerang col-
laboration [13] allows to delineate the region in the (S1M , S1A) plane towards which the
observations point.
Those acoustic peaks appear clearly in Fig. 7 where the temperature anisotropies
across the sky have been expanded in spherical harmonics.
<5T (fJ, <1» = L aim Yim (fJ, <1» . (116)

1m
Each coefficient CI of the ordinate corresponds to the quantity lalm l2 aver-

aged over the various values of m at fixed multipole number t. It represents
the average value over the sky of the temperature difference between two
directions separated by an angle () ~ 200 0 / l. The size of the first acous-

tic peak is directly related to the horizon scale ()H. The latter is mostly
sensitive to the geometry of the universe and its curvature nK .
0.0 I
1.0
OPEN
0.5
0.0
CLOSED -0.5
I
0.0 0.5 1.0 1.5
QA
Figure 9. This triangle shows how the supernovae searches, the analysis of the eMB
fluctuations and the measurement of the gas fraction in X-ray clusters can constrain the
cosmological parameters OM and OA [14] .
In a spherical geometry, the sum of the inner angles of a triangle exceeds

7r and two parallel lines may cross each other. Conversely, defocusing takes
place in a hyperbolic space where an extended region sustains a smaller

angle than in the fiat and spherical cases. The value of () H is related to
452 PIERRE SALATI
~h = nM + nA through
(117)
The measurement of the position of the first acoustic peak is crucial since
it provides the sum of the cosmological parameters nT as shown in Fig. 8.
Supernovae searches and the analysis of the temperature fluctuations in the
cosmic background radiation will eventually allow to determine precisely
the cosmological parameters nM and nA. The uncertainties are still fairly
large but the observations seem to indicate that nT = 1. If true, such a
result would lead to nM= 1/3 and nA = 2/3.
2. The astronomical dark matter puzzle.
In 1933, the Swiss astronomer Zwicky determined the dispersion velocity

of the galaxies inside the Coma cluster. That cluster lies 50 Mpc away in
the direction of the Coma Berenices constellation. It contains thousands
of large galaxies as well as a host of smaller members. Zwicky found that
the dynamical mass of Coma was 10 to 100 times larger than what may
be inferred from the luminosity of the system. The dark matter puzzle was
born. The presence of large amounts of unseen material around galaxies
and inside clusters has since then been confirmed by various independent
methods. An invisible component dominates in many cases the dynamics
of the systems in which it is found. Determining its nature has challenged
quite a few astronomers and the problem is still awaiting for its solution. I
will review here three ways to probe for the presence of that hidden matter.
The kinematic methods based on the rotation curves of spirals and on the
velocity dispersion of ellipticals are presented in section 2.1. The determi-
nation of the mass of galactic clusters from their X-ray emission profiles is
discussed in section 2.2. Finally, the gravitational lensing of distant sources
by a cluster on the line of sight is examined in section 2.3 together with
the determination of the surface mass density of the corresponding lens.
2.1. GALACTIC ROTATION CURVES AND KINEMATIC METHODS.
Spiral galaxies spin on themselves. Their rotation velocity Vc may be mea-

sured through the Doppler shift of the 21 em emission line of neutral hy-
drogen. Quite surprisingly, the rotation curves of many spirals remain flat
far beyond the visible disk. This observation points towards the presence
around these galaxies of extended haloes of dark material. A typical spiral
comprises therefore a luminous disk surrounded by a massive halo. The
latter dominates the dynamics of the system at large distances. The total
mass M is actually related to the rotation velocity Vc through
Vc 2 G M(r)
= (118)
Should the mass be confined inside the disk, the rotation curve would de-
crease like 1/ Vr far from the center. This is not what is observed. In many
cases, the rotation velocity is constant so that the mass must increase lin-
early with distance as
Vc 2
M(r) = O r . (119)
The total galaxian mass MT is therefore proportional to Vc 2 R where R is

the radius of the spiral with its halo. If the spin axis makes an angle ~ with
respect to the line of sight, the spectral lines from the galaxy are broadened
as a result of disk rotation by an amount
~V ~V(O)
-----'---'- sin~ , (120)
c c
where V(O) ~ 2 Vc. The widening of the lines is the largest if the galaxy is
seen edge-on.
Problem nO 22 - Level [1]: The fact that the luminosity of a spiral L is

proportional to its total mass MT is a well-known property of disk systems.
Let us assume furthermore that the average mass density of spirals is also a
constant that does not depend on the object. Show that L varies as V(0)3.
The distance modulus of a source is defined as the difference between the
apparent and absolute magnitudes /-l = m-M. Show that the corresponding
distance is given by
(121 )
The apparent magnitude m of a spiral is corrected for its inclination. Should
the galaxy be edge-on, its brightness would be m(O). Derive the relation
/-l = 7.5lg10~V(0) + m(O) + C . (122)
'fully and Fisher [15] have shown that an empirical relation holds between
the luminosity L and the velocity broadening ~ V (0) of spirals
L ex ~ V(0)2.5±O.3 . (123)
454 PIERRE SALATI
They have inferred a very useful relation that allows to determine the dis-
tance of remote spirals from the rotation broadening of their lines
/-L = 3.5 + 6.25 Ig10 ~ V(O) + m(O) , (124)
where ~V(O) is expressed in km s-l. The Tully-Fisher relation is an effi-

cient distance indicator. A similar relation holds for elliptical galaxies that
appear as oblate spheroids more or less flattened. The Faber-Jackson law
relates the total luminosity L to the velocity dispersion (J of the stars inside
the system
(J ~ 220 kms-
L
1 { Lo '
}1/4 (125)
where Lo = 1010 h- 2 L 0 . The mass-to-light ratio of spirals lies in the range

MIL", 10 - 30 M01 L 0 . For ellipticals it is also quite large with MIL", 20
M01 L 0 · Ellipticals are therefore surrounded by massive haloes just like
spirals. A popular model for those haloes is the isothermal sphere where
the stellar distribution function takes the form
(126)
where E = v 2 /2 + (f) is the mechanic energy per unit mass. Velocities

are isotropically distributed as f depends only on v 2 . The one-dimensional
velocity dispersion is denoted by (J.
Problem nO 23 - Level [1]: Compute the average of the velocity squared

and show that
(127)
Problem nO 24 - Level [1] : Show that if the stellar distribution function

f depends only on E, it is a stationary solution of the Vlasov equation
(128)
where the acceleration of a particle derives from the gravitational potential

through
a= 9 = - V . (129)
Demonstrate eventually that
o. (130)
We will assume that the halo is a sphere. Integrating out the velocity if
leads to the density profile
P (r) = Pc exp { - 7(r)} ' (131)
where Pc is the central value and where the radius r is the only relevant
parameter. The gravitational potential follows the Poisson equation
1 d { 2 d}
~ = r2 dr r dr = 4 1f G p( r ) . (132)
Problem nO 25 - Level [2] : We look for a solution to the Poisson equation

of the form p = A rQ . Compute the parameters A and a to show that the
density of an isothermal sphere varies according to
(j2 1
p(r) = 21fG r2 (133)
Integrating the density p leads to the linear variation of the mass

(j2 _ Vc2
M(r) = 2(jr = a r . (134)
Inside an isothermal sphere, the rotation curve is therefore fiat with a ve-
locity Vc = J2 (j. This seems to be the case for the Milky Way disk even
if observing its rotation is hard. The circular speed at the Earth is
(135)
for a galactocentric distance rev = 8.5 kpc. If our galactic halo was responsi-
ble alone for the rotation velocity measured at the solar circle, its density in
the solar neighborhood would be Pev '" 0.012 Mev pc- 3 . In particle physics
units, this translates 4 into'" 0.47 GeV cm- 3 . We cannot neglect the con-
tribution arising from the disk itself, especially in the inner part of the
4Notice that a density of 1 Mo pc- 3 is equal to 38.3 GeV cm- 3 .
456 PIERRE SALATI
Milky Way. A more correct value for P0 is rv 0.3 GeV cm- 3 . The mass of
the halo far from the center - and therefore the total mass of the Milky
Way - varies like
M(r) ~ 1.12 x 10 11 M0 {lO~Pc} . (136)
We infer a mass of rv 1012 M0 in the inner 100 kpc. The size of the halo is
actually difficult to estimate as the uncertainties are large.
Problem nO 26 - Level [1]: The radius of the halo is denoted by Rh.

Compute the gravitational potential at distance r from the center and show
that the escape velocity in the solar neighborhood may be expressed as
Vesc 2 = 2 Vc 2 { 1 + In -Rh} . (137)

r0
No star is seen in our vicinity with a velocity in excess of rv 500 to 600 km

s-l. This cut-off is interpreted as the local escape velocity from the Milky
Way. Show that Rh is comprised between 40 and 130 kpc.
To summarize, the kinematic methods are based on the assumption that

the Virial theorem applies. The kinetic and gravitational energies are equal
up to some numerical factor. If the system under scrutiny has typical size
R and its components have a velocity dispersion of a, its mass obtains from
the relation
(138)
This is actually the method which Zwicky followed when he estimated the
dynamical mass contained in the Coma cluster.
2.2. HOT GAS AND X- RAY CLUSTERS.
The giant elliptical galaxy M87lies some 15 Mpc away from us at the center
of the Virgo cluster. Its stars have a velocity dispersion a rv 500 km S-1.
If we model the distribution of matter inside M87 with an isothermal halo,
we end up with a mass an order of magnitude larger than for the Milky
Way at same radius
M(r) 1.2 x 1012 M0 {r}

-k-
10 pc
. (139)
From the Faber-Jackson law, we estimate a luminosity of'" 6.3 x 10 11 L 0 ·

The mass-to-light ratio inside the inner 100 kpc reaches a somewhat typical
value of", 18 M0/ L0 - see above. If we now assume that the gas inside and
around M87 is also virialized, the one- dimensional velocity dispersion of the
protons should also be given by CY. The corresponding thermal agitation is
associated to a temperature of
(140)
The gas that lies at the center of the Virgo cluster should be therefore
extremely hot as its temperature reaches", 3 keV. It should be completely
ionized and the plasma should emit X-rays through the bremsstrahlung
process
(141)
As a matter of fact, the X-ray emission from M87 has actually been de-
tected and studied by X -ray observatories such as the Einstein, Exosat or
Ginga satellites. As the gas falls in the deep potential well of the Virgo
cluster, its temperature rises as the result of the conversion of the gravi-
tational energy into specific heat. Note that a temperature of a few keV
is only possible if baryons have completely decoupled from the microwave
background. In particular, attention should be paid to the possibility that
thermal contact between matter and radiation could be re-established if
the intra-clus;ter gas is reionized too early. Remember that the electron
temperature - and therefore the gas temperature - is prevented from in-
creasing too much by the photon thermostat as long as the relaxation time
scale t q is smaller than the age of the universe. This thermal lock no longer
operates if ionization takes place after a redshift of '" 10. The formation
of clusters and the concomitant heating of their inner gas have therefore
taken place recently.
The intensity of the X- ray emission depends on the square of the elec-
tron density n~ and on the plasma temperature T. Analysis of the X-ray
spectrum allows to disentangle T from ne' Some assumptions are neces-
sary to translate the X- ray surface brightness into a spatial distribution
of luminosity. In general, spherical symmetry is assumed in order to de-
rive the electron density ne(r) and the temperature T(r) profiles. The
next step amounts to postulate that the intra-cluster gas is in hydrostatic
equilibrium. This may not be the case close to the center where the den-
sity is so large that the gas has already cooled down efficiently through
bremsstrahlung emission. That gas falls as it fails to be pressure supported
hence the cooling flows that have been observed in the central regions of
clusters. Let us nevertheless assume that hydrostatic equilibrium holds in
458 PIERRE SALATI
the outer regions. The pressure P and the mass density p of the gas follow
the relation
dP
dr = -pg(r) , (142)
where 9 denotes the gravitational field of the whole system.
Problem nO 27 - Level [2] : The density of the intra-cluster gas is n = ne +

nion while its pressure is P = n k T. Translate the condition for hydrostatic
equilibrium into
dlnn dInT 1 GflM(r)
(143)
dlnr + dlnr = kT r
where fl is the mean molecular weight. Assuming a constant chemical com-
position, derive the mass profile M(r)
M (r ) = _ k T r {d In ne + dIn T } . (144)
G fl dlnr dlnr
Problem nO 28 - Level [1] : The inner cluster gas has a primordial compo-
sition of 75% of Hand 25% of 4 He in mass. Show that its mean molecular
weight is fl = (16/27) mB c:::: 0.6 mB whereas the mass per electron is
fle=(8/7)mB.
It is therefore possible to derive the total mass M as a function of the radius

r from the density and temperature profiles of the intra-cluster hot gas.
In the case of M87, these distributions have been determined by Tsai [16J
from measurements of the X-ray surface brightness
(145)
and
T(r) = Too {
r }Q2 . (146)
r + a2
The best fit to the data is obtained for no = 4.31 x 10- 2 cm- 3 , al = 6.63
kpc, al = 0.49, a3 = 0.869, Too = 8.35 X 107 K, a2 = 4.58 x 105 kpc and
a2 = 0.114.
Problem nO 29 - Level [2]: Using relation (144) and the density and
temperature profiles obtained by Tsai, derive the total mass distribution in
M87
(147)
The plasma temperature is '" 3.2 X 107 K at 100 kpc from the center, in
good agreement with our estimate (140). Show that the whole mass of M87
increases linearly with r like
M(r) ::::: 1.3 x 1012 M8 {r}

--
lOkpc
. (148)
It is remarkable that the kinematic determination (139) of the mass con-

tained in M87 agrees within 10% with the result (148) based on the analysis
of the X-ray surface brightness of that source.
Problem nO 30 - Level [2] : Compute the average electron density at radius

r and show that
(149)
Derive the distribution of the hot gas throughout M87
r }1.65
Mgas ::::: 5.3 x 10 9 M8 { - k - (150)
10 pc
The hot gas contained in the central region of the Virgo cluster contributes
a fraction
{
r }O.65
fgas '" 4 X 10- 3 10 kpc (151)
to the total mass of the system. Its relative abundance increases with radius.
The stellar component that dominates the inner parts of M87 becomes on
the contrary negligible with distance. At a radius of 500 kpc which we
may define as the boundary of the cluster, the gas amounts to '" 5 %. Some
clusters are more massive than Virgo and have deeper potential wells. Their
inner gas is heated up to higher temperatures that may reach up to 10 keV.
As these clusters are more bound by gravitation, the hotest part of their gas
460 PIERRE SALATI
has not leaked away. The corresponding gas contribution fgas is therefore
slightly larger than what has been derived for the central part of Virgo
and its dominant galaxy M87. A survey of 45 X-ray clusters [17] indicates
actually that the gas fraction is remarkably stable over the sample, at least
for a temperature in excess of '" 5 ke V
fgas = {0.07 ± 0.002} h- 3 / 2 . (152)
A Hubble constant of h = 0.65 translates into a gas fraction of 13 %. The

existence of that plateau is remarkable. It suggests that fgas is represen-
tative of the cosmological abundance of baryons with respect to the bulk
of the non-relativistic matter. Clusters of galaxies are the largest known
structures and should be typical of the cosmic average. Following this line
of reasoning will lead us to another estimate of the cosmological parameter
OM·
Problem nO 31 - Level [1]: Relate the cosmological baryon abundance

OB to the baryon-to-photon ratio 7]10
OB h 2 = 3.64 X 10- 3 7]10 . (153)
The primordial abundance of deuterium may be measured inside a few ab-

sorbing clouds located on the line of sight to distant sources. BurIes and
Tytler have analyzed four high-redshift hydrogen systems seen in absorp-
tion against distant QSOs. In the case of the object QSO 1937-1009 for
instance, the intervening cloud is located at a redshift of z = 3.572. BurIes
and Tytler derive [18] a deuterium abundance of
(D / H)p = (3.4 ± 0.3) x 10- 5 . (154)
That element is a sensitive baryometer and allows to probe for OB. The
previous measurement translates into 7]10 ~ 5.5 and
OB h 2 = 0.02 ± 0.002 . (155)
If we now assume that the gas fraction inside clusters is indicative of the
ratio OB/OM on cosmological scales, we are led to the conclusion that the
mean matter abundance in the universe is given by
OM = fOB = (0.29 ± 0.04) h- 1/ 2 (156)

gas
We readily derive a value of

OM ::: 0.36 ± 0.05 (157)
where a Hubble constant h = 0.65 has been taken. This determination is
based on the X-ray surface brightness of clusters and on the assumption
that these systems are representative of the cosmological scales. It is in fair
agreement with the result derived from the supernovae searches in the case
of a flat universe
OM ::: 0.28 ± 0.13 , (158)
Two independent methods lead therefore to the same value of the mat-
ter abundance. The mere existence of such a concordance - the so-called
cosmic concordance - is quite exciting. It strengthens the conclusion that
OM "-' 1/3 and makes a strong case for the existence of vacuum energy or
a cosmological constant.
2.3. GRAVITATIONAL LENSING.
I will introduce here a new way to probe for the matter contained inside
galaxies and clusters based on the deviation of light by gravitating bodies.
I will show how a mass distribution may act as a lens on distant sources
lying on the line of sight beyond the deflecting system. This field has been
developing in the last decade in what is now a mature domain of extra-
galactic astronomy. I will also use the main results of this section later on
when I discuss the gravitational microlensing. To commence, I will focuse
on matter distributions that generate a weak gravitational field. By weak I
mean that the potential wells of the systems at stake are associated to non-
relativistic virial velocities v < < c. In this limit, the stress-energy tensor
of such a distribution is
(159)
where p denotes the mass density and UJ.L = dxJ.L / dT is the 4-velocity. The
pressure is negligible as it comes as a v 2 correction. The system comprises
particles whose velocity field is described by iJ(P, t). Depending on the scale,
those particles may be stars or galaxies or the putative neutralinos that will
be discussed in section 3.3. In the weak field limit, the gravitational field
may be expanded around the Minkowski metric
(160)
where h is small with respect to unity. The perturbation hJ.L1/ is related to

the mass density p (P, t) and to the velocity field iJ(P, t) through
hoo = 2 and the vector potential V which the system
generates are defined as
cI>(M,t) = _G r p(P,t') dTp

c 2 } distribution r
V(M, t)
= _G r pv(P, t') dTp (162)
c3 } distribution r
The potentials (162) correspond actually to the retarded solution a la Lien-

ard et Wiechert of the linearized Einstein's equations (13). Each element
dTp generates at time t' a contribution to cI> and V that propagates from
P to M at the speed of light.
Problem nO 32 - Level [3] : Starting from the Riemann-Christoffel curva-

ture tensor (11), show that the Ricci tensor may be expressed as
ga{3
2 {8a 8{3 g/-LV - 8a8v g/-L{3 - 8/-L8{3 gav + 8/-L8v ga{3}
(163)
Linearize the Ricci tensor in the weak field limit to demonstrate that the
Einstein's equations of gravitation become
(164)
The source S/-LV is computed as if the metric was flat. This leads to
S/-LV = T/-Lv - ~ 1]/-LV T\ = P { U/-L Uv - ~ 1]/-LV} , (165)
where O(v 2 ) terms are neglected.
Problem nO 33 - Level [2] : Impose the condition corresponding to the

gauge of the harmonic coordinates
(166)
to get the propagation equation for the gravitational potentials hJ.Lv
DhJ.Lv = - 16 7r G SJ.LV . (167)
Solve it and derive the retarded solutions (162).
The velocities of the matter distribution under scrutiny are non-relativistic.

The crossing time by light or by a gravitational field of such a system is
negligible with respect to its characteristic evolution time scale. Our Milky
Way for instance rotates in 200 million years whereas a photon crosses
it entirely in less than 2 x 105 years. Light propagates basically instanta-
neously throughout the system. We can therefore approximate t' by t in
relations (162). From now on, we will also neglect the time dependence of
the potentials <1> and V as the source which generates them changes slowly
in time and behaves essentially as if it was static. The propagation of light
is associated to a vanishing interval of proper time dT so that a photon
travels over a distance dO M = eds in the time interval
dt ::: (1 - 2<1» ds + 4V· dOM . (168)
Notice that the vectors have here their usual Euclidean meaning. In partic-
ular, e is a unit vector whereas the distance ds between two points M and
M' next to each other follows the conventional definition
ds 2 = 8mn dxmdx n = dOM·

- -
dOM . (169)
In general relativity, light rays and free fall trajectories for that matter are
understood as the geodesics of a manifold curved by gravitation. We will
adopt here an alternate view. We will consider that space- time is actually
flat. Photons travel along bent lines because the medium throughout which
they propagate is optically inhomogeneous. Gravitation is accounted for
effectively by an optical index of refraction n(M) that varies with position.
Problem nO 34 - Level [2] : In a medium where the speed of light is v,

the index of refraction is defined as n = c/v. Derive the effective index n
generated by our non-relativistic distribution of matter
n=1-2<1>+4V·e. (170)
We would like to determine the trajectories followed by light in such a

medium. Let us consider for that purpose such a path M(s) which starts
464 PIERRE SALATI
from point A and ends at point B . The distance along that line - the
curvilinear abscissa - is denoted by s. Let us now perturb that trajectory
M(s) into the path P(s) so that !vip = 80M = £(s). The departure and
arrival points are still the same. According to the Fermat principle, the
travel time tAB from A to B is stationary with respect to the perturbation
t of the trajectories.
Problem nO 35 - Level [3]: Compute the variation 8tAB of the travel

time from A to B between the real and the perturbed paths. Expand that
variation up to first order in the trajectory modification t. You may first
show that
_ d, -, dV
atAB = iA{B - ds€· { - (ne) - \1n
ds
+ 4 --} +
ds
4dse· 8V , (171)
where n' = 1 - 2<p. Derive the eikonal equation

de - _ -
(1 - 2<P) ds = - 2 \1 ~ <P + 4 e 1\ rot V (172)
The unit vector e is tangential to the light path. The gradient operator '\7
may be decomposed into a tangential part e (e· \1) and a perpendicular
component '\7 ~ with respect to e.
The motion of light under the action of gravity is reminiscent of the tra-
jectory followed by a charged particle that would interact with an electro-
magnetic field. The first term in the right-hand side of Eq. (172) is the
analogue of the electric force. The second term may be interpreted as a
gravito- magnetic interaction where the role of the magnetic field is now
played by rot V = '\7 /\ V. Notice that the vector potential V is actually
defined in the same way as its magnetic counterpart A. Because the source
of the gravitational field in which light is deflected is non-relativistic, the
gravito- magnetic contribution e/\ rot V to the eikonal equation is negligi-
ble with respect to the electric-like term '\7 ~ <P. In the weak field limit, the
scalar potential <P < < 1 and the eikonal equation simplifies into
de -
ds = - 2 \1 ~ <P . (173)
We would like to establish now the relation between the positions of a

source and its images as seen in the plane of a gravitational deflector. In
order to derive the so-called lens equation, we first consider a light of ray
that propagates along a straight line in the absence of gravity. Its tangential
vector is denoted by ein. A point- like mass m is located at point 0 and
bends inwards the light trajectory in the plane defined by 0 and ein. After
the deflection, the light points towards the new direction characterized by
the unit vector e out::: ein'
Problem nO 36 - Level [2] : Integrate the eikonal equation (173) along

the unperturbed trajectory and show that the total deflection due to the
point- like mass m amounts to
4Gm {
ex = ein - e out (174)
~~2
The vector of least approach { points from the mass m towards the light
trajectory with respect to which it is perpendicular so that (. ein = O.
Show that a light of ray that grazes the surface of the Sun is deflected by
an angle
u"' = 4GM28 ::: 175. arcsec. (175)
R8 c
The deflecting body is a static distribution of matter located around the ori-
gin 0 of our coordinate system. The mass density is specified by p(x, y, z).
If unperturbed, the light would travel along the vertical axis Oz and ein =
ez . Each point P of the deflector contributes a mass element dm
p(P) dr(P) .
Problem nO 37 - Level [1] : By summing the effect of each mass element

dm, compute the total deviation of the light ray as it passes through the
lens
ex = ein - eout = :~ / / dxdy 'L,(x,y) ~~ . (176)
The vector (is horizontal and points from P towards the axis Oz
(177)
The deflection ex depends only on the surface mass density
'L,(x, y) = ! p(x, y, z) dz . (178)

466 PIERRE SALATI
The latter corresponds to the projection of the mass on the plane of the sky
as seen by an observer sitting close to Oz. Relation (176) may be transposed
to the general case where the light trajectory is still vertical but now crosses
the plane (0 x y) at point A defined by 011 = a
(179)
Each point M of that plane is located by the vector [ = OM and asso-

ciated to the mass element ~([) d2 { As long as the deflection along the
unperturbed trajectory is a good approximation to the real deviation, the
deflecting body behaves exactly as if it was a thin lens whose mass would
be compressed along the plane (0 x y) perpendicular to the line of sight.
In the two cases presented in these lectures, this so-called thin lens ap-
proximation holds even if the system at stake is a sphere. Let us assume in
addition that an observer is located on the vertical axis Oz - the lens axis
- at point T such that ZT = Dd > O. The observer is collecting the light
from a distant star E located in the vicinity of Oz beyond the deflector
at position ZE = -Dds < O. In the absence of gravitational deflection, the
light ray would propagate along the straight line connecting E to T and
would cross the lens plane (0 x y) at point I. The real trajectory is actually
bent by the deflector. It crosses now the lens plane at point A. Notice that
the incident direction ein points from E to A and differs slightly from the
outgoing vector eout that connects A to T. We note
OJ = if position of the source

011 a position of its image(s) (180)
as projected on the lens plane.
Problem nO 38 - Level [2] : In the paraxial limit where the distances of E,

I and A to the lens axis Oz are small with respect to Dd and Dds, derive
the basic equation that maps a source I to its image(s) A
if = a - D~~d &(a) (181)
The distance between the observer and the source is Ds = Dd + Dds.

A galactic cluster such as Abbel2218 shows an intricate system of giant
arcs and more modest arclets. In order to discuss the gravitational lensing
that such a system generates, we start from the fact that the potential well
is generally dominated by the central galaxy and is well-described by an
isothermal halo whose properties have already been discussed in sections 2.1
and 2.2. Matter is isotropically distributed around the cluster center O.
Problem nO 39 - Level [2] : Evaluate the surface mass density ~ (f)of an

isothermal sphere whose velocity dispersion is (7 and whose density profile
is given by relation (133)
(72
~ (f) (182)
2G~
The deflection ii of a light ray that crosses the system obtains from rela-
tions (179) and (182)
(a - f)
(183)
Ila-fW
A direct computation of the previous integral may seem tedious. Fortu-
nately, the calculation of the gravitational distortion ii considerably sim-
plifies when one realizes that the problem at stake has a well-known elec-
trostatic analogue. Let us actually consider a vertical wire whose charge
per unit length is A. The wire crosses the plane (0 x y) at point M located
by the vector [ = 0 AI. The electric field generated in that plane at point
A such that a = OA may be expressed as
A (a - [)
lIa - [W
dE{M -+ A} = (184)
2 IT 100
Problem nO 40 - Level [3]: The gravitational distortion ii is the ana-

logue of the field generated by an ensemble of vertical wires that carry the
f
electric charge A = Pel ( ) d2 per unit length. This corresponds to a 3D
distribution whose density is
Pel (f) (185)

468 PIERRE SALATI
Use Gauss theorem to compute straightforwardly the integral (183)
(186)
The lens equation associated to our naive isothermal sphere relates a source
if to its images a as seen projected on the sky onto the plane of the deflector
(187)
The Einstein radius of the cluster is defined as
(188)
If the source is a remote galaxy or quasar whose light is deflected by a

cluster of galaxies that lies halfway on the line of sight, the Einstein radius
is seen by the observer under an angle
(189)
that does not depend on the distance in the case of a flat space. For typical
velocity dispersions'" 500 - 1000 km s-l, we infer an angle ()E '" 3 - 15
arcmin. This is well within reach of the modern telescopes, especially of the
4 and 8 meters class. The source I and its images A as given by the lensing
cluster are aligned with the center 0 of the potential well. Once projected
on the radial line joining 0 to I, they satisfy the equation
a
a - rJ = RE ~ . (190)
A graphical solution of that equation allows to distinguish three regimes .

• Close to the cluster center, the distortion effect may be critical and gives
rise to multiple images that appear as giant arcs. Our toy model shows
that if the source lies within the Einstein ring - rJ < RE - three images are
generated : Al is just at the center of the lens (al = 0), A2 is still inside the
Einstein ring on the opposite side with respect to I (a2 = rJ - RE) while the
last image A3 lies outside that critical region at position a3 = rJ+ RE. Notice
that the only effect of the lens is to shift the radial positions of the images
with respect to the sources. It preserves their polar angles as seen from the
center of symmetry o. A small disk is therefore transformed into a very
elongated arc that lies along the Einstein ring. An interesting application of
the existence of giant arcs is the possibility to measure the Hubble constant
Ho by monitoring the radio emission from the multiple images of the same
radio-source. Each image is associated to a different travel time to the
Earth. Measuring the delay between those various replicas gives directly
access to the physical size of the source and hence to its distance .
• When the source crosses the Einstein ring, its images Al and A2 merge
together and vanish as soon as 1] > RE. The last image A3 is left alone.
As long as the background source is close to the region of critical lensing,
the distortion remains significant in spite of the absence of multiple images.
This is the so- called arclet regime .
• When the source is far from the Einstein ring, the distortion gets fainter
as a3 = 1] + RE ~ 1]. The background galaxies have nevertheless a tendency
to be flattened along the radial direction towards the center of the lensing
system. This effect is hard to observe on a single object as any particu-
lar galaxy does appear more or less elongated depending on its intrinsic
ellipticity or orientation. It is measurable on a field of sources through
the apparent correlation of their orientations. This so- called weak lensing
allows to determine the surface mass density of the lens as projected on
the sky and probes the dark matter distribution lying inside a cluster of
galaxies.
3. Machos versus Neutralinos.
In the previous section, I have described three methods to scrutinize the

distribution of matter inside galaxies and clusters. They all point towards
the presence of large amounts of unseen material that dominate the inner
gravity of these systems. The evidences for dark matter on galactic and
cosmological scales are not disputable. The puzzle lies in the fact that
the very nature of that invisible mass has not yet been resolved. A host
of conventional and exotic candidates has flourished. We will review here
two of them. Dark baryons will be discussed in section 3.1 whereas the
neutralinos of the supersymmetric theories will be presented in the two last
parts 3.2 and 3.3.
3.1. BARYONIC DARK MATTER.
We do expect the presence of dark baryons. Primordial nucleosynthesis

leads on the one hand to a total baryon abundance of
0.05 ± 0.005 (h = 0.65) , (191 )

470 PIERRE SALATI
whereas on the other hand the luminous component of the universe - stars
and galaxies - contributes a smaller density of
nv ~ 0.01 . (192)
• Dark baryons may be in the form of cold gas. The column density of
neutral hydrogen (HI) is obtained from surveys at radio wavelengths of the
galactic 21 cm hyperfine emission. Molecular hydrogen H2 is not detected
directly. It clusters in clouds and its distribution is inferred from the tracer
molecule carbon monoxide (CO). The transition between the rotation levels
J = 1 -t 0 of the latter, detected at 2.6 mm, plays the role of the hyperfine
transition for neutral hydrogen. The conversion coefficient X that translates
the integrated CO line blackbody temperature Weo into a column density
of molecular hydrogen N (H 2 ) has been determined empirically in our galaxy
X = N(H2) = 2.3 x 1020 moleculescm- 2 / (Kkms- 1 ) , (193)

Weo
and could be significantly underestimated. The possibility arises therefore
that the halo of the Milky Way is made of clumps of very cold H2 that would
basically go undetected. This scenario may nevertheless be constrained for
clouds lying in the vicinity of the galactic ridge, a region where cosmic
rays are accelerated and propagate. These high-energy particles interact
with the interstellar gas to produce in particular photons. Should they be
impacted by cosmic rays, the putative clouds of cold and therefore dark hy-
drogen would eventually shine at ,-ray energies. This would result into an
additional component to the gamma-ray diffuse emission from the galaxy.
As no such component is seen, we may exclude [19] the presence of cold
clouds in the neighborhood of the Milky Way within a region extending
over '" 3 kpc above and beneath its disk .
• Another possibility lies in the existence of dark stars or Jupiter-like ob-
jects. Below a limit of", 0.08 M 8 , the inner temperature of a star is too low
for hydrogen to fuse into helium. The resulting brown dwarf cools down to
an inert state where the crushing gravity is supported by the Fermi pressure
of the degenerate electrons. The infrared emission of the star is quite faint,
hence the difficulties to discern its presence. A decade ago, Paczynski [20]
reasoned that a brown dwarf - or any point-like object for that matter -
would deviate the light from a distant background star in quite the same
way as the deflectors discussed in section 2.3. Let us once again consider
an observer T receiving the light from a source E in the same configura-
tion as before. A deflecting body with mass M lies at the center 0 of our
coordinate system. The corresponding surface mass density along the lens
plane (0 x y) is
(194)
It generates the gravitational deflection
_ 4GM a
a = - 2- -2 (195)
c a
The lens equation associated to this so-called Schwarzschild deflector read-
ilyobtains
-_-{I
ry-a R~}
--2
a
' (196)
where the Einstein radius is now defined by
R _ 4GM DdsDd
E - Ds (197)
c2
As seen on the sky in projection on the lens plane, the deflector 0, the
background source I and its images A are all aligned. The images Al and
A2 are respectively located outside and within the Einstein ring.
Problem nO 41 - Level [2] : Show that a Schwarzschild lens gives two

images Al and A2 of a source I located at
(198)
When their alignment is perfect, one could conclude intuitively that the
background star is concealed behind the deflector which would block its
light. As a matter of fact, the source appears as a circle centered on the
lens - the so-called Einstein ring - and its luminosity is enhanced as will be
shown subsequently. In the case of a star located in the Large Magellanic
Cloud - a satellite galaxy of our Milky Way - the angular radius of the
Einstein ring generated by a 1 M0 deflector lying halfway would be ,...., 0.4
milliarcsec. The power resolution of the Hubble Space Telescope reaches 0.1
arcsec and is three orders of magnitude above the sensitivity limit required
to observe such a gravitational mirage. That is why this phenomenon is
called gravitational micro lensing as the multiple images that it generates
cannot be distinguished.
Notice that the optical configuration associated to gravitational lensing
is nearly symmetrical with respect to the deflector plane. Exchanging the
source E with the telescope T would require to rigorously redefine the lens
axis as the straight line EO. To a very high accuracy, the latter may be
472 PIERRE SALATI
approximated by 0 z so that our lens plane is still (0 x y). As seen from

the source, the telescope would normally appear on the sky as point I. Its
mirror or radio dish would have a projected area onto the deflector plane of
d2 if. Gravitational lensing replaces that collecting area by an image whose
surface has become d2 ii. As the source radiates isotropically, the amount of
energy received by the telescope is enhanced in the presence of the lens by
a factor of
(199)
where J denotes the Jacobian associated to the mapping of if into ii. The
surface luminosities of the source and of its images as seen now by the
observer E are identical.
Problem nO 42 - Level [2] : Show that the Schwarzschild lens amplifies

the brightness of the source by an amount
A = Ial dal I + Ia2 da21 (200)

ry dry ry dry
The reduced distance between the source I and the deflector 0 - as seen
in the lens plane - is defined as u = ry/ RE. Compute the amplification
factor A as a function of u and derive the basic relation of gravitational
microlensing
u 2 +2
A= (201)
uvu 2 +4 '
Whenever a source I lies within the Einstein ring of a point-like deflector,

its brightness is enhanced by a factor of rv 1.34 which corresponds to a
shift of rv 0.3 magnitude. This effect is potentially detectable but attention
must be paid to disentangle such a micro lensing event from a variable star.
Fortunately, the luminosity enhancement induced by the close angular en-
counter of a background star with a lens located on the line of sight has a
very peculiar signature .
• The light amplification is achromatic because the effective gravitational
optical index does not depend on the frequency.
• In projection on the sky, the background star moves along a straight line
that passes in the vicinity of the lens. The microlensing is symmetrical in
time with respect to the moment where the angular distance u between the
source and the deflector is minimal and where peak luminosity is reached.
• Finally, gravitational microlensings are very rare events as we shall see be-
low. A background star can be lensed only once as the corresponding prob-
ability is vanishingly small. Should a given source exhibit several changes
of its brightness, it would be immediately classified as an intrinsic variable
and disregarded as a potential candidate for a microlensing event.
Paczynski's original idea was to use gravitational microlensing to probe
for the presence of dark stars inside the galactic halo. The static probability
T that any given background source is amplified by more than 0.3 magnitude
amounts to compute the number of deflectors N 1ens that lie along the line
of sight within an Einstein radius from the latter. This probability is called
the optical depth for gravitational microlensing and may be expressed as
T == N 1ens = j dx nRk nlens(X) . (202)
Problem nO 43 - Level [1] : Show that the optical depth T depends only
on the mass density Plens of the deflectors along the line of sight x
T = 4nGj dx
~ Plens(X)
{X(Ds-X)}
Ds . (203)
The Milky Way halo is well-described by an isothermal sphere and we may

assume that the deflector density is given by 5
a2 1
Plens '" 2 7r G x2 . (204)
Note that we have mistaken the distance x along the line of sight with the
galactocentric radius r. In spite of this gross oversimplification, we may
nevertheless derive a rough estimate for the optical depth against a remote
source located at distance D s
T '"
Vc22
- {Ds
In- + -r0 - 1} . (205)
c r0 Ds
The Small (SMC) and Large (LMC) Magellanic Clouds contain millions of
giant stars that may be monitored separately. They are far enough so that
a large part of the galactic halo may be probed. In the case of the LMC for
which Ds ~ 52 kpc, we derive an optical depth of
T '" 5.2 X 10- 7 , (206)

5S ee also relation (133).
474 PIERRE SALATI
where a value of rev = 8.5 kpc has been assumed. More precise estimates
give a comparable value of 4.5 x 10- 7 for the LMC and of 6.8 x 10- 7
for the SMC. The duration of a microlensing event depends on the relative
transverse velocity v 1.. - as projected on the lens plane - between the source
and the deflector
t event = -RE (207)
vJ..
A detailed investigation shows that should the halo be made of brown

dwarves or Jupiter-like planets, one gravitational microlensing is expected
each year per million of stars. Such events have actually been observed
in the direction of the clouds but there are fewer of them than predicted.
Only two SMC events have been reported and there are approximately 20
LMC events in all. In the latter case, the combined EROS and MACHO
measurements lead to an optical depth of
(208)
A recent analysis points towards an even smaller value. The Milky Way
halo comprises only a fraction of point-like deflectors. From the lack of
short duration microlensing events, a firm upper bound of 10% has been
f'V
established at the 95% CL on the contribution of light compact objects with

mass lying in the range from 5 x 10- 7 up to 2 X 10- 3 Mev. Two events were
induced by a binary lens. In that case, the distance could be measured and
put the deflectors in the clouds themselves. We already know that the SMC
is elongated along the line of sight. For the LMC, a self-lensing explanation
is a little more involved. In spite of its irregular shape, this satellite galaxy
is actually a disk which appears nearly face-on with an inclination of its
rotation axis of ~ ~ 27°.
Problem nO 44 - Level [2] : A flat infinite galactic disk is made of a

single population of stars whose velocity dispersion - in the perpendicular
direction Oz - is denoted by a. We assume that the stellar distribution
function is well accounted for by relation (126). The stellar density along
the vertical axis is therefore given by
p(z) = exp {_ .p(z)} (209)

Po a2
Solve the Poisson equation
= 47rGp (210)
and show that the mass distribution across the galactic disk may be ex-
pressed as
p(z)
= cosh- 2 {_z_} , (211)
Po V2a
where the typical scale a is defined as
(J
a = J47rGpo
(212)
For a distant observer, the fraction (Ds - x) IDs that appears in rela-
tion (203) may be approximated by unity provided that x denotes the
distance between the lens and the source.
Problem nO 45 - Level [3] : Compute the optical depth that corresponds

to the self-lensing of the galactic disk against its stars located on the plane
z = O. Notice that the def1ector- to-source distance x amounts to zl cos E
and derive the expression
T (213)
Gould [21] has computed the complete optical depth against all the stellar
sources spread inside such a disk. His exact result is close to our previous
estimate
(214)
The vertical velocity dispersion of red-giant stars is rv 20 km S-l. This

leads to a self-lensing optical depth against the LMC of rv 1.1 X 10- 8 - an
order of magnitude below the observations. More recently, a bias has been
found [22, 23] in the vertical velocity dispersions of the LMC which may
have been underestimated by a factor of rv 3. If so, the measured value of
T ~ 1 - 2 X 10- 7 may be entirely explained by the lensing of the LMC
stars on themselves. The nature of the Milky Way halo remains therefore
an open question.
3.2. NEUTRALINO SYNTHESIS IN THE EARLY UNIVERSE.
The cosmological abundance of non-relativistic matter amounts to

+0.08 +0.05
OM (OK = 0) = 0.28 _ 0 0. 9 (stat) _ 0.04 (syst) (215)
476 PIERRE SALATI
whereas the bulk of the baryons contributes

OB h 2 = 0.02 ± 0.002 (216)
Room is left for the presence of non-baryonic candidates to the astro-

nomical dark matter. The infatuation for neutral and weakly-interacting
elementary particles lies in the fact that should such species exist, they
would have been processed in the early universe in such a way that their
relic abundance would be today in the ballpark of OM. The possibility that
a large fraction of the matter in the universe is non-baryonic is also in
agreement with the scenarios of galaxy formation. As already mentionned
in section 1.3, baryons decouple from the radiation at the end of recom-
bination. In addition to this photon-baryon mixture, a new distribution
of basically weakly-interacting and pressure-less matter should also ex-
ist. Relation (113) implies that the density fluctuations of that cold dark
matter (CDM) collapse at all scales and form the potential wells inside
which baryons fall to form galaxies and clusters at the end of recombina-
tion. In this section, I will discuss the particular case of neutrinos as the
prototype of a more general class of particles that comprises the so-called
neutralinos. I will first establish the relation between the temperature and
the age of the universe. I will show that the plasma cools down rapidly as
it expands. The primordial behaviour of elementary particles is therefore
quite different from what may be guessed from our daily life. Instead of
thermalizing and reaching an equilibrium with the surrounding matter, the
various constituents of the primeval gas decouple from each other. This de-
parture from thermal equilibrium may occur while the species still behave
as a pure radiation or after they have started to be non-relativistic. I will
discuss separately these two possibilities .
• The dynamics of the early universe.
At early times, matter is completely dissociated into its basic constituents.
The universe is filled with pure radiation and its energy density may be
expressed as
(217)
For simplicity, I will work from now on in a system of units where 1i = k =

c = 1. The effective number of degrees of freedom geff(T) of the primordial
plasma - what Gamow called the Ylem - depends on the populations that
are ultra-relativistic. This happens whenever the temperature exceeds the
mass of a given species so that the coefficient geff(T) may be written as a
sum over the relevant boson gB and fermion gp spin states
(218)
Problem nO 46 - Level [1] : Show that for a temperature of 1 Me V, the

energy density of the primordial plasma is given by geff = 43/8. Three
families of light neutrinos, each with two helicity states, will be assumed
here.
The entropy density of the plasma may be expressed as
(219)
Entropy conservation implies that the product a a 3 is conserved. As the de-

pendence of geff with the temperature is weak, we infer that T decreases as
a-I. The expansion of the universe is therefore determined by the evolution
equation
T
T
= ~ = {8 G
a 3
1f
geff
(T) 1f2
15
T4} 1/2 .
(220)
Problem nO 47 - Level [2] : Show that expression (220) may be written as
(221)
If the variations of the coefficient geff with respect to the temperature are
neglected, show that the age of the universe, i. e., the time t since the bang,
is related to the temperature T by
1 { 32 1f3 G } 1/2
T2 ~ 45 geff(T) t (222)
You need to take, as the initial condition of the problem, an infinite tem-
perature at time t = O.
In order to evaluate numerically the previous expression, the Planck mass

needs to be defined. That mass - or energy - is associated to Newton's
constant of gravity G = 6.67 X 10- 11 m 3 kg- 1 S-2. It may be multiplied by
the appropriate cocktail of fundamental constants 'Ii and c in order to derive
478 PIERRE SALATI
a quantity whose dimension is an energy. The Planck mass is therefore

defined by
Mp = Jh~5 ~ 1.96 X 109 J ~ 1.22 X 1022 MeV . (223)
Because an inverse MeV corresponds to a duration of hl1 MeV, z.e., nu-

merically to 6.6 x 10- 22 seconds, relation (222) translates into
t ~ 1.7 second {I MeV}2 . (224)

Vgeff(T) T
During primordial nucleosynthesis for instance, the temperature decreases
from 1 MeV down to 0.1 MeV, i.e., from 10 to 1 billion Kelvins. The age
of the universe at that time lies in the range between one second and three
minutes .
• Light neutrinos and the thermal quenching.
At low energy, the weak force proceeds through a point- like interaction
connecting four fermionic currents. In the case of the proton-neutron trans-
mutation for instance, the effective Lagrangian may be expressed as
At low energy, the relevant coupling parameter of weak interactions is the

Fermi constant
GF = 1.166 X 10- 5 GeV- 2 . (226)
The vector 9v and axial 9A couplings of the neutron-proton weak current
have been experimentally determined. Their numerical values are respec-
tively 1 (vector) and 1.26 (axial). The neutrino coupling is a pure chiral
interaction with projection only on the left states through the operator
PL = (1 - 15) /2. Chirality becomes identical to helicity in the limit where
the mass of the particle is negligible with respect to its energy. In our case,
the plasma temperature is typically 1MeV. Neutrinos interact with the
f<V
surrounding particles with typical cross section
< <7V > f<V GF2 T2 f<V 1.6 x 10- 33 cm3 s-1 { T }2. (227)
1 MeV
The collision rate of neutrinos with the species of the surrounding medium
may be expressed as
(228)
Problem nO 48 - Level [2] : Compute the density nA of an ultra-relativistic

population of particles A. In this regime, the temperature T is much higher
than the mass M of the species at stake - electrons or muons or pions or
_ 9A T3 (3) { I (Boson) (229)

nA - 71"2 ( 3/4 (Fermion)
where ((3) = 1.20205. In order to restore the conventional units, show that
the previous expression needs to be multiplied by
(nkJ 3 = 83.22 K- 3 cm- 3 . (230)
In that regime, remarkably enough, the ratio nA / T3 is constant. For pho-

tons, it is approximately equal to 20 K-3 cm- 3 . For neutrinos with two
spin states, it is rv 15 K-3 cm- 3 .
The neutrino population gets thermalised with the rest of the plasma
through the numerous collisions which the particles have together. If the
interactions are efficient enough, energy is democratically shared among
the various species. The neutrino temperature Tv is therefore the same as
for the rest of the plasma. The collision rate r c describes the rapidity with
which a fluctuation of the neutrino temperature is erased. It quantifies the
propensity of neutrinos to get thermalised. Because T decreases as the in-
verse of the scale factor a, the rate of its evolution is set by the Hubble
expansion factor
(231)
Note that the neutrino collision rate r c is more temperature sensitive and
decreases faster than the expansion rate H. At high temperature, r c is
larger than H. This results into the thermalisation of the neutrinos when
the temperature T is larger than the critical value Td := 1.5 MeV for which
r =
H
c 1 := 0.3 { Td
1 MeV
}3 . (232)
Below Td, the probability for a neutrino to collide with another particle
during a typical expansion time 1/ H has become less than unity. The
primordial plasma is completely transparent to neutrinos which behave as
a fossil population without any further interactions. When T rv 0.5 MeV,
480 PIERRE SALATI
electrons and positrons start to annihilate with each other into photons.
Heat is released. Because now neutrinos no longer interact, they do not
benefit from that energy production. The photon bath alone is reheated.
As a result, the photon temperature T slightly increases with respect to
Tv. The variation of the T lTv ratio may be estimated by noticing that
the entropy of the radiation, i.e., the mixture of photons, electrons and
positrons, remains constant.
Problem nO 49 - Level [1] : A volume with typical size ()( Tv -1 is called a

covolume as its expansion follows the growth of the universe. Compute the
entropy of the above-mentionned radiation before and after the electron-
positron annihilations. Show that the ratio {T lTv} 3 increases by a factor
of 11/4.
The increase of the photon- to-neutrino temperature ratio translates today

into a neutrino temperature of Tvo ~ 1.95 K. Note that after the neutrino
decoupling, Tv is more a measure of the dilution than a real temperature.
The density of neutrinos decreases as a- 3 . Relation (229) may still be used
provided that Tv ()( a-I. We infer a relic density of
Pv
°= 3({3)
2 7[2
T03
v
M '" 110k V
- e cm
-3 {~}
1 ke V ' (233)
where M stands for the neutrino mass. That result is to be compared to

the closure density
P~ = 1.879 X 10- 29 g cm- 3 h 2 ~ 10.6 keY cm- 3 h 2 . (234)
The neutrino relic density p2 may be conveniently expressed in units of the

closure density to yield
(235)
The larger M, the larger the relic abundance Ov h2 . Requiring that Ov ::;
OM '" 0.36 and setting h = 0.65 leads to the bound on the sum of the
neutrino masses
Lmv ::; 15 eV . (236)
v
Note that such a limit seems to be satisfied. The mass of the electron
neutrino cannot exceed a fraction of an eV. The deficiency of the elec-
tron neutrino flux produced at the solar core as well as the deficiency of
the muon neutrino flux produced by the spallation of high-energy cosmic
rays impinging on the top of the atmosphere point towards the existence
of oscillations among the various neutrino flavours. In particular, such a
solution in the case of the atmospheric neutrino puzzle favours a value of
f:j.m 2 = 3.2 x 10- 3 eV 2 for the vJ.L H V T oscillation .
• Non-relativistic decoupling.
If the neutrino mass exceeds the decoupling temperature of Td ~ 1.5 MeV,
neutrinos may substantially annihilate before freezing out. As an illustra-
tion, we focus here on the specific case of a heavy particle A whose mass
is M = 2 Ge V. The analysis which I will present now is generic to an en-
tire class of massive and weakly-interacting species. In particular, it may
be readily transposed to the supersymmetric relics of section 3.3. At high
temperature, for T 2: 2 GeV, heavy neutrinos are in chemical equilibrium.
They steadily annihilate into light If fermion pairs while the reverse pro-
cess is also very active. The annihilation-production reaction
A+A~I+f (237)
is in equilibrium and the A's density relaxes towards its equilibrium value.
Below 2 GeV, A and A annihilate and the population becomes non- relativistic.
Problem nO 50 - Level [1] : As long as the chemical reaction (237) is in

equilibrium, show that the A's density is given by
a
n A = 9 A T3 e - a ( 2 7r
)3/2 ' (238)
where a denotes the ratio MIT. Because entropy is conserved, the scale
factor of the universe evolves as T- 1 . The ratio MIT plays therefore the
role of such a scale factor a{t) that increases while the universe expands.
As the temperature decreases, the A's are significantly depleted by anni-

hilation and their density drops down. The antiparticles A with which A's
annihilate also become rare. At T ~ 100 MeV, the density nA is so low,
particles and antiparticles are so much depleted that reaction (237) ceases
to be in equilibrium. The probability for an A to encounter an antipartner
becomes less than unity per typical expansion time H- 1 . Annihilations stop
under the combined action of the dilution due to the expansion of space,
and of the severe depletion of the A population which has occurred between
T = 2 GeV and T = 100 MeV. Below 100 MeV, annihilations are inhib-
ited and the density of particles per covolume remains constant. Around
Td ~ 1.5 MeV, heavy neutrinos stop colliding with the other species and
482 PIERRE SALATI
become mere fossils of the early stages of the universe. If they are stable,
they pervade the intergalactic medium until the present epoch, and may
even contribute a significant fraction to the mass density OM.
Assuming that there are as many particles A than antiparticles A, the
density nA evolves according to the differential equation
dnA 2
dt = -3 H nA - < O'an V > nA 2 + < O'an V > n~ (239)
100
UR fA
fA3quilibrium
!"I
I
-E
t"';l
0.01 2 GeV Dirac Neutrino
~
.5
"8~ 0.0001
l!::!
0..
><
o
Decoupling at aF = 18.06
\
\
le-06 \
\
\
\
\
\
\
\
\
,
\
le-08 \
\
\
\
\
\
,
\
\
1e-1O ,
\
\
,
\
\
,,
\
\
1e-12 '--_""---'--'-...L.-I............LJ....._--'----'...........--'-..........r...£..L_ _' - -........-'-L-I-J....A..J..J

1 10 100 1000
Mass to Temperature Ratio a = Mff
Figure 10. The codensity fA == nAIT 3 - solid - is presented as a function of the ratio
a == MIT, for a 2 GeV neutrino. At large temperatures, it approaches its ultra-relativistic
value of 15 K- 3 cm- 3 - UR label on the vertical axis. The codensity follows the ther-
"V
modynamical equilibrium f~ - dashed line - down to the critical point where decoupling
occurs. In this example, the freeze-out temperature corresponds to aF 18. Then, the "V
codensity is unable to follow f~ which rapidly decreases. After decoupling, fA smoothly

reaches its present value.
where the first expression on the right-hand side refers to the dilution result-
ing from the expansion. The second term accounts for the A annihilations
while the last expression describes the back-creation of the AA pairs from
light fermions and assumes detailed balance. The density n~ corresponds to
thermodynamical equilibrium. In the non-relativistic regime at stake here,
it is given by relation (238). In terms of the codensity I A = nAIT3, the
evolution equation simplifies into
dlA 310 2
dt + «O"anv>nA)IA =<O"anv>T A . (240)
Because entropy is conserved, the temperature varies as the inverse of the

scale factor a. The codensity I A corresponds therefore to the number of
particles inside a volume ex: a 3 that expands with the expanding universe.
In order to solve the differential equation (240), two typical time scales
need to be defined.
1) When equilibrium is reached, i. e., when IA = I~, the time derivative
diAl dt ~ O. The characteristic time scale of the relaxation of I A towards
its kinetic equilibrium value I~ is related to the annihilation rate
(241)
2) The time scale of the variations of the equilibrium I~ itself may be

expressed as
(242)
in the non-relativistic regime.

As is clear in Fig. 10 and 11, two stages may be distinguished during the
decoupling .
• At high temperature, as long as Trel < T eq , I A has plenty of time to
relax towards the equilibrium I~ which evolves at a much slower pace. As
a consequence, the annihilation reaction (237) is in chemical equilibrium so
that I A = I~· In Fig. 10, the solid line I A cannot be distinguished from
the dashed curve I~. As is clear from Fig. 11, as T decreases and the ratio
MIT increases, relaxation becomes progressively less efficient until the ratio
Trez/Teq eventually reaches unity where the decoupling from equilibrium
occurs .
• Below the freeze-out temperature Tp, the relaxation time Trel exceeds
Teq. Whilst I~ drops down and vanishes, I A still decreases a little bit. The
annihilation cross section of a pair of heavy neutrinos is
- - G~ 2
O"an V (AA -+ If) = - M NA (243)
211"
484 PIERRE SALATI
le+06
Evolution of the equili~ri~'m
Relaxation toward the equi!.i6rium .......... /
100000
10000
::0
s:::
0
C,)
II)
Vl
0
.... 1000
...."">/
C,)
§
Vl
... "
:,
II) ~~
"; 100 "...
"
C,)
Vl
II)
," :...
E
f=: IO
///-,'-" !
";
C,)
",,,,,
'6.
:>. /'
E-<
-'
,,,,"
,,-,'
/ Decoupling
0.1
0.01
10 100 1000
Mass to Temperature Ratio a =Mrr
Figure 11. The typical time scale Trel with which the annihilation reaction relaxes
towards its chemical equilibrium is plotted - dotted curve - as a function of the ratio
a = MIT . The chemical equilibrium itself evolves with the timescale Te q featured by the
solid line. The age tu of the universe corresponds to the dashed curve.
Problem nO 51 - Level [3] : Show that the decoupling condition Trel = Teq
translates into
_ 3 J5 2
JCiF eaF 3
- (211")4 GF M Mp NA geff
-1/2
~ 0.7xlO
7 { M }3 N A
1 GeV .jg;ff
(244)
That relation corresponds to a non-relativistic decoupling since the ratio
M /T is equal at freeze-out to aF 20. For a 2 Ge V neutrino, the effective
I"V
number of annihilation channels is N A 14 whilst the effective number I"V
of degrees of freedom is geff = 57/8 for a temperature of TF 100 MeV. I"V
Deduce from above that aF ~ 18.06.

In Fig. 11, the freeze-out is indicated by the arrow labelled Decoupling.

As soon as the ratio Trel/Teq exceeds unity, the codensity fA - solid line of
Fig. 10) - de couples from the equilibrium value f~ - dashed curve. From
decoupling until now, fA has been slightly decreasing according to the re-
lation
(245)
which, in terms of the parameter x = l/a, simplifies into
df A
--2
_
-
{g5
--3 geff
-1/2
< aan V > M Mp dx (246)
fA 87f
Problem nO 52 - Level [2]: Integrate that differential equation from

XF = l/aF where the codensity decouples from equilibrium, i.e., where
fA(aF) = f~(aF) = fF, until the present epoch x = 0, where fA reaches
its asymptotic value f.f Y. Show that
(247)
At present, the co density of a stable 2 Ge V neutrino is f ;sy '"

4 X 10- 9
which translates - after multiplication by a factor of {k/n c} = 83.22 K- 3 cm- 3
3
- into'" 3.3 x 10- 7 cm- 3 K- 3. Note the generic dependence of f.fY on the
annihilation cross section : the larger < a an V >, the stronger the annihila-
tion and subsequent depletion, the lower the relic abundance f;:SY.
Problem nO 53 - Level [3] Translate relation (247) into the relic abun-
dance
° _ 1.23 keV cm
PA - -3
geff
1/2
aF {TJ.}3
7K {3 X 10- 27 cm3 s-l } . (248)
2. < aanV >
The antiparticles A - that exist if the species under scrutiny is a Dirac

fermion - contribute the same density. The effective number of degrees of
freedom at decoupling is denoted by geff. Today's temperature TJ. of the A
distribution takes into account the series of annihilations-reheatings that
took place since decoupling.
486 PIERRE SALATI
10
Majorana
Dirac
,, Maximum of 1
C'l
,,
< ,,
..c:: ,,
,
-co
='
c
I
I
I
I
~I I
\",
E
o 0.1
'" ,
v
u
"
c
~
"0
C
"'" """
;:I
.0
~ Minimum of II40 -';"-"-"-"-"-,,-,-..>0,.,----
0.Q1
""
"""
,,
,
""""
0.001
10 100
Neutrino Mass [GeV]
Figure 12. The abundance Q v h 2 of heavy remnant neutrinos is plotted as a function

of their mass M for Majorana (solid curve) and Dirac (dashed line) fermions [25). The
parameter Q v is defined as the ratio of the neutrino relic density p~ to the critical density
pS. Note that Q v h 2 decreases approximately like M- 2 • Below"" 1/40, it is so low that
it cannot account for the dark matter around galaxies and inside clusters.
Problem nO 54 - Level [3]: Show that the relic abundance of heavy

neutrinos may be expressed as
(249)
Derive the relic density
(250)
to which both neutrinos and antineutrinos contribute. Take geff = 57/8 and
N A = 14 to infer the numerical value
n 2 {1.8GeV}2 (251)
Hvh rv M .
Compare to Fig. 12 where the annihilation cross sections of heavy Majo-

mna (solid curve) and Dimc (dashed line) neutrinos have been numerically
computed. Any comments?
Notice finally that Ov h 2 is peaked around M rv 5 MeV and is fairly well

described by expression (235) for M < 5 MeV whilst relation (251) holds
above. A heavy Dirac neutrino with mass in the range rv 2 - 10 Ge V would
contribute a significant fraction to the closure density pg
today should it
exist and be stable. For Majorana neutrinos, the relevant mass range is
between 3 and 20 GeV as their annihilations are suppressed with respect to
their Dirac partners. As a matter of fact, the LEP results point towards the
existence of only three light neutrino species. Those were already known to
exist in association with the electron, the muon and the tau leptons. The
existing limits on their masses are discussed elsewhere in the proceedings of
the Cargese School. A fourth family of heavy neutrinos may well exist after
all but such a species should not contribute to the ZO decay width, i. e., it
should not be produced in ZO decays. Its mass must be larger than M Zo /2 rv
46 GeV. Should such a particle exist and be stable, its remnants would not
contribute significantly to the closure mass density of the universe as is
clear in Fig. 12. A heavy neutrino cannot be a good dark matter candidate
[26]. The LEP accelerator has played the role of a dark matter telescope in
disregarding heavy neutrinos as the explanation of the astronomical dark
matter puzzle.
3.3. DESPERATELY SEEKING SUSY.
Fabio Zwirner has explained elsewhere in these series of Cargese lectures the
motivations for introducing supersymmetry in high-energy physics. Suffice
it to say that if supersymmetry holds, there should be an entire new zoo of
elementary particles that differ from the conventional species by a half-unit
of spin. If R-parity is a conserved quantum number, the lightest element
among the supersymmetric bestiary is absolutely stable 6. It furthermore
corresponds to a neutral particle in an extended region of the supersymmet-
ric parameter space. The so-called neutralino is typically a massive particle
with mass in the range 30 GeV - 1 TeV whose interactions are weak, on
6That particle is called the Lightest Supersymmetric Particle - LSP.
488 PIERRE SALATI
the order of < GanV > '" 3 X 10- 27 cm 3 s-l. We can readily estimate its relic
density from relation (248).
Problem nO 55 - Level [1] : Setting geff '" 40, aF '" 20 and Tl '" 1 K
and remembering that supersymmetric neutralinos are Majorana fermions,
show that they contribute today a mass density of
3 x 10- 27 cm 3 s-1 }
Ox h2 '" { (252)
< GanV >
In most cases, the neutralino X turns out to be the lowest-mas superposition

of the photino 1', the zino Z and the two higgsino fip and fig states - see
Zwirner's contribution
(253)
Depending on whether ar + a§ is larger or smaller than a~ + a~, the LSP

is called a gaugino or a higgsino. If the astronomical dark matter is made
of neutralinos, we should float in a gas of such particles. Because they are
weakly interacting, detecting the presence of the neutralinos is extremely
difficult. I will respectively discuss the direct and the indirect signatures of
these supersymmetric candidates for the dark matter .
• Direct searches. In spite of their weak interactions, the putative neu-
tralinos should nevertheless undergo a few collisions with ordinary matter.
In such events, a heavy particle with velocity Vx '" J3/2 VC '" 270 km S-1
impinges on a nucleus at rest that recoils as a result of the energy transfer
ER .
Problem nO 56 - Level [1] : Show that the recoil energy of the nucleus
may be expressed as a function of the center of mass scattering angle 0*
J.L2
ER = (1 - cos 0*) - v; , (254)
mN
where J.L is the reduced mass of the neutralino- nucleus system

1 1 1
= + (255)
The neutralino velocities required to generate a specified recoil energy

ER lie in the range extending from the minimum value of
(256)
up to the escape velocity Vrnax from the Milky Way as seen in the Earth
reference frame.
10- 5
, \
~ /
'\
/
10- 6 /
\ /
\ /
\
\\ "- ./'
/'
E
..,
~
.E.
10- 7 \ '" \
§:; "
""
\ ~
~B ? /
1~ "- '------ ---:::-
b
10 - 8
10~
Figure 13. Upper bounds on the scalar neutralino-nuc1eon cross section. The dot- dashed
line features the limit obtained by combining the various Ge experiments [27] . The
solid [28] and the long-dashed [29] curves denote the bounds set by the NaI detectors.
The dotted and short-dashed limits are based on data utilizing Xe [30] and Te02 [31].
490 PIERRE SALATI
Problem nO 57 - Level [2] : Show that the energy spectrum of the recoiled
events generated by neutralinos may be expressed as
dR
dER
= NT Px r
mx }Vmin dER
da Ilvll f (v) d 3 v ,
max (257)
where NT is the number of target nuclei per unit mass of the detector. The
local density of neutralinos is Px while their velocity distribution f (v) is
generally assumed to be given by the exponential function (126).
\
. '-
.\
'- .
10'
m)( (GeV)
Figure 14 . Upper bounds on the spin-dependent neutralino-nucleon cross section. The

dot-dashed line features also the limit obtained by combining the various Ge experi-
ments [27J. The solid [28J and dashed [29J curves are derived from NaI experiments. The
dotted line corresponds to data from a Xe detector [30J.
The differential cross section da / dE R depends on the impinging velocity v

and on the recoil energy ER. It may be conveniently split into a coherent
part and a spin dependent component. The former arises from the inter-
actions of the neutralino onto the target quarks through the exchange of
scalar particles such as the Higgs bosons and the squarks. That is why this
term is also called the scalar contribution. As it is coherent, it varies as
A2 with the nuclear mass number A. Elements such as 73Ge or 129Xe have
high atomic numbers and their cross sections with neutralinos is large. If
Z-exchange dominates, the cross section depends now on the spin of the nu-
cleus. That nuclear spin cannot reach as large values as A. The limits set on
the cross section are therefore much weaker when the spin-dependent inter-
actions dominate as is clear from Fig. 13 and 14 [27]. Detecting the galactic
neutralinos through their energy deposition inside terrestrial instruments
is a real challenge. There is a considerable background which needs to be
removed. That is why the various ongoing experiments are underground to
get rid of cosmic rays. They also operate at very low temperature around
a few tens of milliK. They tend finally to be based on the double detection
that allows to cross examine the events in the light of two different chan-
nels such as the heat versus the ionization or the scintillation. In order to
discriminate the neutralino events from the mere local radioactivity, one
may notice that the signal dR/dER depends on the particle distribution
f (v) where v denotes the velocity as seen in the Earth frame whereas the
Maxwellian exponential (126) is expressed in the galactic frame. This leads
to an annual modulation of the neutralino direct detection rate as the Earth
moves around the Sun .
• Astrophysical signatures.
Should neutralinos pervade the halo of the Milky Way, their mutual an-
nihilations would yield several indirect signatures that are potentially de-
tectable on Earth as additional components to the various cosmic radiations
- W+W- , ... ---t "{,p,- D- ,e+ & v I s .
X + X ---t qq, (258)
Detection of these annihilation products relies on several experimental tech-
niques which allow to probe different regions in the (mx ' (1) plane. Searches
for antiprotons, antideuterons and positrons are performed by balloon and
satellite-borne devices. Because the flux depends on the square of the neu-
tralino density n x , the limit which may be set on the annihilation cross
section scales approximately as mx 2. That type of search is mostly sen-
sitive to low neutralino masses. High-energy photons are detected both
by air Cerenkov telescopes (ACT) and by satellite-borne instruments. The
neutralino annihilation rate, and hence the gamma-ray signal, also scale
as mx -2. Because of the background in which that signal is swamped, the
corresponding limit on the annihilation cross section approximately scales
492 PIERRE SALATI
as the mass m x ' This channel compares therefore with the direct searches,
a technique that is sensitive to the neutralino density nx = PDM/m x and
that sets the same kind of limit on CT. On the contrary, the neutrino channel
offers the opportunity to probe for large values of m x ' The corresponding
limit on the annihilation cross section scales this time as mx -1. The heav-
ier the neutralino, the more stringent the bound. This channel is therefore
complementary to the other searches.
Clumpiness.
Dark matter may have partially collapsed in domains of larger than aver-
age density. A fraction f of the neutralinos may actually be in the form
of clumps [32, 33, 34] inside which the annihilation rate, and hence the
corresponding indirect signals, are increased by a factor of
(259)
A clumpy distribution of neutralinos leads therefore to the enhancement

C = f x 8 of the various annihilation radiations.
- In galactic haloes, the clumps should follow the CDM density fluctuations
(260)
that are related to the comoving wave vector k through the spectrum [35]
(261)
with a = 1.71 xl, fJ = 9 X l1.5 and'Y = l2 where l = (f2h2)-1. Normalization

to CT8 = 0.8 gives A = 2.82 X 106 Mpc 4 when f2 = 1 and h = 0.5. In a
restricted wavelength range, it is approximated by a power law
P(k) oc kn . (262)
The power spectrum P(k) of density fluctuations behaves as k- 3 on small

scales, i.e., for structures typically lighter than Mi rv 108 M 0 . As regards a
possible clumpy structure of the halo around the Milky Way, the relevant
mass range extends from Mi rv 108 M0 up to Ms rv 1012 M 0 . The corre-
sponding spectral index n goes from -2.6 to -2.1. When a numerical value
is needed, we will take the effective value of n = -2.36. Structures smaller
than Mi turn out to all have the same density. They contribute identically
to the clumpiness factor C. There is no larger structure than the halo itself
whose mass Ms reaches 1012 M0 in the inner 100 kpc. Because the comov-
ing wave vector k scales as M- 1 / 3 , neutralino density fluctuations depend
on both the scale M and the redshift z as

Jp ex: (1 + Z)-l M-(n+3)/6 (263)
p
The redshift factor (1 + z) -1 is typical of the t 2 / 3 growth of density fluctua-
tions in a flat matter-dominated universe - see our discussion of section 1.3
and relation (114). Notice that small scale perturbations, for which n = -3,
all become non-linear at the same time. Their subsequent collapse leads to
virialized structures whose densities have been enhanced by a factor of ,. . .,
180 with respect to the epoch of formation, when Jp/ p reached unity. Small
scale dark matter clumps all have therefore the same density today. The
formation redshift of larger structures behaves as
(1 + ZF) ex: M-(n+3)/6 , (264)
so that today, neutralino clumps with mass above,....., 108 M0 have a density
p(M) ex: 180 (1 + ZF)3 ex: M-(n+3)/2 . (265)
The density p(Ms) of the largest possible clump should be comparable to
the average dark matter density PDM in the galactic halo.
Problem nO 58 - Level [1] : Assume that a fraction f of the latter is

exclusively made of small clumps whose mass does not exceed Mi ,. . ., 108
M0 and derive
= {P(Mi)} = (Ms) (n+3l/2 (266)
C f p(Ms) f Mi .
Relation (266) translates numerically into C = 20 f. For a discussion of

the case where the clumps span the mass range from Mi up to M s , see
Ref. [36]. Depending on the fraction f, the enhancement of the indirect
signals at stake does not exceed a factor of,....., 20.
- Neutralinos may cross the Sun or the Earth. If they interact in their inte-
riors and lose enough energy to get trapped, they eventually concentrate at
the cores of those celestial bodies and built up a dense distribution. That
is why the center of the Sun and, in a lesser extent the center of the Earth,
are bright neutrino sources.
- Neutralinos may also be clumped at the galactic center. If they are dis-
tributed so as to built up an isothermal halo, their density profile is given
by relation (133) where a core radius a r-v 1 - 4 kpc is now included
p(r) ex: (r2 + a2) -1 . (267)

494 PIERRE SALATI
We have adopted a solar neighborhood value of P0 ,...., 0.3 GeV cm- 3 . N-

body simulations of structure formation show that a central cusp should
appear at the centers of galaxies such as the Milky Way. The formation of
that cusp drags down some of the cold matter at the core of the system
so that the neutralino density profile should be quite peaked in the central
region [37, 38, 39]. According to [37], the neutralino density diverges as r- 1
r- 1
(268)
p(r) 0:: (1 + rja)2 '
with P0 ,...., 0.3 GeV cm- 3 and a,...., 25 kpc. That central cusp could actually
be a bright source of high-energy photons potentially observable by the
next generation of ACT's.
- Last but not least is the effect of the massive black hole that sits at
the center of the Milky Way. Monitoring of the stars in that central region
[40] shows evidence for the presence in the inner,...., 0.1 pc of a celestial
body whose mass reaches MBH = 2.6 X 106 M 0 . When collapsing, that
black hole should have also dragged down some of the neutralinos floating
around. Starting from an already cuspy distribution with Pinitial 0:: r-'"Y
where T = 0 - 2, P. Gondolo and J. Silk [41] have shown that if the black
hole formation is slow enough as to be adiabatic, neutralinos condense into
a central spike whose density profile is given by
P final 0:: r -'"Ysp . (269)
The index TSP varies from 1.4 for an initial isothermal distribution to 2.5
in the case of an initial cusp with T = 2. For simplicity, we will assume
here that the neutralino central spike amounts to a fraction ), of the cen-
tral black hole so that MDMspike = ), MBH. Annihilations in that core
should not proceed too efficiently under the penalty of erasing the neu-
tralino condensation. Assuming a cosmological density relic of 0xh2 = 0.1
translates - according to relation (252) - into an annihilation cross section
of,...., 3 x 10- 26 cm 3 s-1.
Problem nO 59 - Level [1] : Because the annihilation rate in that central

spike cannot exceed the inverse of the age of the galaxy - ,...., 10 Gyrs - show
that the neutralino density is bounded by
(270)
where mlOO = mx/lOO GeV and derive the corresponding spike radius
1/3 -1/3
a spike "'0.13pcA m lOO . (271)
High-energy gamma rays.

Neutralino annihilations lead both to a continuous spectrum of high-energy
photons as well as to monochromatic gamma-ray lines through the box-
mediated direct production
x + X -+ , +, & , + ZO . (272)
Problem nO 60 - Level [2] : Derive the neutralino-annihilation ,-ray flux

at the Earth
n"susy _ ~ (O'v) N,
'1'
,
-
47l' m~
J
P2 ds.
los X
(273)
The flux (273) may be split into an astrophysical piece and a particle
physics part. The former term merely amounts to the integral along the
line of sight of the dark matter density squared Px 2 . For a neutralino dis-
tribution with total mass M and typical size R, it is approximately given
by M2 R- 5 . Its bench mark value is '" 0.1 M0 2 pc- 5 in the case of the
Milky Way, for R = 100 kpc. The particle physics piece depends on the
annihilation cross section (O'v) averaged over the distribution function of
the neutralino velocities. That distribution function is typical of the system
under scrutiny. For the halo of our galaxy, the one-dimensional dispersion
velocity is 0' = Vc/V2 rv 160 km/s whereas for the giant elliptical galaxy
M87 at the center of the Virgo cluster, it reaches", 500 km/s. In relation
(273), N, denotes the total number of continuous or monochromatic pho-
tons that are produced in a single annihilation. In the latter case, N, = 2
for the two-photon reaction while N, = 1 for the ,-Zo channel. Because
the ,-ray signal from annihilating dark matter is faint, the most promising
instruments are the atmospheric Cerenkov telescopes. When a high-energy
photon impinges on the top of the atmosphere, it generates an electromag-
netic shower that spreads over a significant area as it reaches the ground.
The initial ,-ray energy is degraded into many optical photons that are
potentially detectable by an array of telescopes. Analysis of the shape of
that shower allows in principle to reconstruct the direction and the energy
of the primary high-energy photon. ACT's get advantage of their large ef-
fective detection area. They suffer however from glaucoma as they can only
496 PIERRE SALATI
monitor small portions of the sky at the same time. They are perfectly
suited for faint and point-like sources such as the ones at stake here, i. e.,
the central spike at the galactic center or the extra-galactic system M87.
Satellite- borne detectors are complementary. They are well suited for large
solid angle surveys but their collecting areas are small. In any case, the
detection is made difficult by the presence of various backgrounds in which
the signal is swamped. The dominant source arises from the cosmic ray
(CR) high- energy electrons that impact on the upper atmosphere and also
generate electromagnetic showers. An ACT cannot distinguish between a
photon and a CR electron- induced shower. The corresponding flux is given
by
q>e = 6.4 X 10- 2 [e- GeV- l cm- 2 S-1 sr- l ] (Ell GeV)-3.3±0.2 . (274)
Hadron- induced showers are more extended on the ground than those of the
electromagnetic type. Stereoscopy is a powerful tool to discriminate hadrons
from electrons and gamma-rays. The CAT experiment, for instance, has
already achieved a rejection factor of one misidentified event over a sample
of 600 showers generated by CR hadrons whose flux at the Earth is
q>had = 1.8 [proton GeV- l cm- 2 s-1 sr- l ] (Ell GeV)-2.75. (275)
For satellite-borne instruments, the only background arises from the , - ray
diffuse emission of interstellar gas that shines under the action of the local
CR protons. The corresponding emissivity may be expressed as
The hydrogen column density in the direction of the galactic center amounts
to NH '" 1.5 X 1023 H cm- 2.
The two-photon line seems to be the most promising reaction to look
for, provided the instrument has a good energy resolution. The resulting
monochromatic photons have an energy E"'( = m x '
Problem nO 61 - Level [2] : Following Ref [42J, translate relation (273)

into
q>~USY = 3.8 X 10- 11 photon cm- 2 s-1 sr- l (O'v)29 m 102 J(u), (277)
where (O'v)29 is the annihilation cross section for the two- photon line pro-
cess expressed in units of 10- 29 cm 3 s-1 while mlO = mxl10 GeV.
-
-27 £. A. Ballz, C. Briol, P.SaJali, J . Silk and R. Tail/el, 1999
___ 10
'~ 30" limit, ll.ElE=O.02
M -28
810
CJ
'--"
~ -29
1\ 10
6V -30
N 10
-31
10
-32
10
-33 2yline
10
x Gaugino-like
-34
Mixed
10 G
+ Higgsino-like
x
10 -35 ~L....L.ifrl:u.L_~..L-;-L.....l......L..L.J.....LL...::::::::::i::::::::L~::u::j~
10 2 10 3 10 4
Neutralino Mass (GeV)
Figure 15. Annihilation rate in the monochromatic channel 2-, line. Each point cor-
responds to a specific supersymmetric configuration. The 3- 0" detection limit is featured
for an optimistic 10 km 2 yr exposure towards M87 together with an energy resolution of
t.E / E = 0.02 [36].
The line of sight integral J is expressed as a function of its bench mark value.
The latter is given by the product r0P~ where a galactocentric distance of
r0 = 8.5 kpc has been assumed together with a halo density of P0 = 0.3
GeV cm- 3 - typical of the solar neighborhood. If a NFW halo profile -
relation (268) - is assumed for the neutralinos clumped at the galactic
center, the line of sight integral becomes (J) rv 103 (1° If)) when averaged
over an angle f) from the center. The next generation of ACT's will typically
have an effective collecting area of 0.1 km 2 , an angular resolution of 0.1°,
a threshold of 50 GeV as well as an energy resolution of ± 15 %.
498 PIERRE SALATI
Problem nO 62 - Level [1] : If the observation of the putative neutralino

clump sitting at the galactic center is performed for 0.1 yr - two or three
months equivalent - show that the collected statistics of the ,-ray line
annihilation photons amounts to
Ny = 1.3 X 105 photons (lTv)29 m1i (:0) , (278)
to be compared to a background of CR electrons of
Ne = 2.9 X 108 electrons m 102 .3 (:0) 2 (279)
The electron background is uniformly distributed on the sky. It exhibits

fluctuations of amplitude ~ which have a smaller chance of being inter-
preted as a signal when the significance S = N'Y / ~ is large. Requiring a
3-lT detection, this method is sensitive down to
(280)
By comparison, a 1 m 2 satellite-borne telescope with an energy resolution

of 1 % would reach a similar sensitivity with
2 (lTv) 'Y'Y >

- 2 X 10- 28 cm 3 s-l m 100 1.13 , (281)
should it monitor for two entire years a region of the sky within 10 from
the galactic center. The small number of collected events is nevertheless a
problem. Whatever the method - ACT or satellite - only the upper fringe
of the supersymmetric configurations in the ((lTv)n , m:\) plane can be
probed.
The situation is quite different if we now assume the existence of a cen-
tral neutralino spike. A precollapse NFW halo profile would be associated
in that case to a present core radius of rv 0.01 pc together with a total
neutralino mass amouting to a fraction ). rv 10- 3 of the central massive
black hole. An ACT would reach a 3-lT detection limit of
2 (lTv) n -> 3 X 10- 34 cm3 s-l m 1 0

-0.15
0, (282)
for a detection cone encompassing a region within 0.10 from the galactic
center. An impressive set of supersymmetric configurations becomes ex-
plorable - see Fig. 15. Should Gondolo and Silk's analysis turn out to be
correct , the galactic center would be a remarkable hot spot on the , - ray
sky.
....I
(a)
M87
C
E
u
L-
a
(b) mx = 1 TeV
I"L-
a(J) Eo = 100 GeV
>,
N
signal
I
E (c)
..:lC
...---..£;
10 2
W
/\
w
..........
(d)
Z '"
10
(e)
-1
10
10
e ( arcmm
. )
Figure 16. The radial profiles of the neutralino-induced signal (solid curves) and of the
various backgrounds (dotted and dashed lines) are plotted as a function of the angular dis-
tance to the source centers. A fiducial model with mx = 1 TeV and {av}cont .N-y = 10- 25
cm 3 S-1 is taken while a threshold of 100 GeV is assumed. The backgrounds are respec-
tively labeled as (a): electronic; (b): hadronic; (c): extragalactic; (d) : M87 and (e): Milky
Way gamma- ray diffuse emissions [36] .
The extra-galactic source M87.

Annihilation photons make it also possible to investigate the presence of
neutralinos in extra-galactic systems. The most promising site is the giant
elliptical galaxy M87 at the center of the Virgo cluster, some 15 Mpc away.
That galaxy is known to contain large amounts of dark matter. Its line of
500 PIERRE SALATI
sight integral of ,...., 10 M0 2 pc- 5 is two orders of magnitude larger than for
the isothermal halo around our Milky Way. Furthermore, M87 extends on
,...., 30 arcmin and appears as a point-like source, well-suited for an ACT
observation. Finally, if a fraction f of the neutralinos that pervade M87 is
in the form of clumps, the annihilation signals are enhanced by a factor of
C ,...., 13 f to 40 f, depending on whether the clumps are smaller or bigger
than Mi ,...., 108 M0 [36]. In addition to the other backgrounds already
discussed above, an extra-galactic component should also be considered as
well as the diffuse emission arising from the in situ spallations of cosmic rays
with the gas inside M87 itself - see Fig. 16. Depending on the fraction f,
the continuum ,-ray signal from M87 is detectable by the next generation
of ACT's for a part of the supersymmetric configurations outlined in the
upper-left panel of Fig. 17. Because low-energy photons are predominantly
produced in neutralino annihilations, the lower the detection threshold, the
better the sensitivity. Finally, even with the annihilation rate enhanced by a
factor of 40, the gamma ray lines are out of reach, at least with the present
and near future instruments.
Antiprotons and antideuterons.
The mutual annihilations of the neutralinos potentially concealed in the
galactic halo could also produce an excess of antimatter particles such as
antiprotons, positrons and even antideuterons. Cosmic ray fluxes are about
to be measured with unprecedented precision both by balloon borne detec-
tors [43] and by space instruments [44]. The various ongoing experiments are
also hunting for traces of antimatter. The search for antinuclei has actually
profound cosmological implications. The discovery of a single antihelium
or anticarbon would be a smoking gun for the presence of antimatter is-
lands nearby. Alternatively, an excess of antiprotons at low energy - below
,...., 1 Ge V - has also been advocated as a potential signal for the putative
supersymmetric dark matter.
As regards the antiproton cosmic radiation, the problem arises once again
from the existence of a background. High-energy cosmic rays, mostly pro-
tons, do produce antiprotons when they interact on the interstellar material
p(CR) + H(ISM) -+ p +X . (283)
Those secondary antiprotons propagate throughout the galaxy in just the

same way as any other species. The erratic structure of the galactic mag-
netic fields results into the diffusive transport of the cosmic rays, both in
the disk itself as well as in the,...., 3 kpc thick layers above and beneath it.
Particles typically spend,...., 5 Myr in the disk where they cross a column
density of,...., 10 g cm- 2 with which they interact or annihilate. During the
remaining 90% of the time, they are confined in the magnetic fields that
have spread out of the disk. If they survive to their diffusion throughout
the Milky Way, they eventually escape into intergalactic space.
-23
~10 ~~~~~--~-'''''nT'---~''"on~
....
'rJ:J -24
r"lS 10
(.J -25
'-';-10
Z..l -26
g10
u
/\ -27
E; 10
V
-28
10
-29
10
x Gaugino-like
o Mixed
-31
10 + Higgsino-Iike
Et~ = 50 GeV
Neutralino Mass (GeV)
Figure 17. Annihilation rates in the continuum channels. The threshold has been set
equal to Eth = 50 Ge V. Considering M87 as the source, the 3-0' detection limits for
exposures of 0.01 km 2 yr are also presented. The region below the heavy solid lines will
not be accessible, even with the next generation of Cerenkov telescopes. The lower solid
line shows the region of accessibility if the annihilation rate is enhanced by a factor of 40
due to dumpiness [36).
Because the secondary antiprotons are not produced at rest, the low-energy
part of their spectrum is expected to be depleted . A '" 10- 20 GeV proton
has actually little chance to produce an antiproton at rest by impinging on
an hydrogen atom of the interstellar medium. Previous calculations of the
seconday p flux at the Earth showed that the spectrum reaches a maxi-
mum for Tp 2 Ge V and significantly decreases below that peak. For just
rv
502 PIERRE SALATI
the same kinematic reasons, the mutual annihilation of neutralinos should

favour on the contrary the production of low-energy primary antiprotons.
The supersymmetric antiproton spectrum is actually flat. There is quite
an excitement to extract from the observations a possible p exotic compo-
nent which would signal the presence of supersymmetric dark matter in the
galaxy.
F. Donalo . N. Fornengo. P. Salali (1999)

103
D AMS/ISSA
10 2
10 1
",........
!=:
'-......
>-Q)
x
0 x
",........
II
C"")
x.x
x •• ,
0.1 XX
~ .~,
.......
0
........., 10- 2
~~~~.x
1"-: x
1-
'S< x
•
/(~
x to::
"
x x
x
~
x
x
x . x
x
Xx x x
t:
IQ x x x x
~ x
~ x
.
.x
10- 3 x x
x
x x
x
x )(
gaugino
10- 4 mixed
0 higgsino
10- 5
0 100 200 300
mx (GeV)
Figure 18. The supersymmetric f> flux has been integrated over the range of IS en-
ergies extending from 0.1 up to 3 GeV In. The resulting yield Nf> of antideuterons [48]
which AMS on board ISS can collect is plotted as a function of the neutralino mass m)(.
Modulation has been considered at solar maximum.
Unfortunately, it has been recently realized [45, 46, 47] that a few processes
add up together to flatten out, at low energy, the spectrum of the conven-
tional secondary antiprotons from which a potential primary component
becomes hard to separate. The dominant energy loss mechanism is the in-
elastic but non-annihilating interactions of antiprotons with the interstellar
protons. The latter are excited towards resonant states and hence absorb
part of the antiproton energy. Ionisation losses marginally contribute. The
low-energy tail of the p spectrum is therefore replenished by the more abun-
dant populations from higher energies. That effect is further strengthened
by solar modulation which also shifts the energy spectrum towards lower
energies. As a result of these effects, the secondary p's are much more
abundant at low energy than previously thought. Disentangling an exotic
supersymmetric contribution from the conventional component of spalla-
tion antiprotons may turn out to be a very difficult task. The antiproton
signal of supersymmetric dark matter is therefore in jeopardy.
Antideuterons , i.e., the nuclei of antideuterium, are a priori free from
such problems [48J. They form when an antiproton and an antineutron
merge together during a spallation reaction or when a neutralino pair anni-
hilates. Both antinucleons must be at rest with respect to each other in order
for fusion to take place successfully. A spallation reaction creates very few
slow antiprotons. The production of low-energy secondary antideuterons is
further suppressed as both antinucleons must be at rest. The energy loss
mechanisms are also much less efficient. With a binding energy of B 2.2 <"'oJ
MeV, the antideuteron is a fragile nucleus. Any interaction, such as an

inelastic but non- annihilating scattering, that would lead in principle to
the 15 energy loss is also associated to an energy transfer that destroys the
antideuteron. On the other hand, supersymmetric D's are mostly manufac-
tured at rest . In neutralino annihilations, antinucleons are predominantly
produced with low energies. This feature is further enhanced by their sub-
sequent fusion into antideuterons, hence a fairly flat spectrum as shown by
F. Donato at this conference. Below a few GeV In, secondary anti deuterons
are suppressed with respect to their supersymmetric partners. That low-
energy suppression is orders of magnitude more effective for antideuterons
than for antiprotons. This makes cosmic-ray antideuterons an alternate
probe for supersymmetric dark matter. Fluxes are nevertheless quite small.
A dozen only of secondary D's should be collected above an interstellar en-
ergy of rv 3 GeV In on board the future space station borne AMS detector.
Less than one secondary event is expected below that value where the bulk
of the primary D's is concentrated. The corresponding yield ND is plotted
as a function of the neutralino mass mx in Fig. 18. A decent amount of
supersymmetric configurations can be probed. In Fig. 19, the sensitivity
limits of the BESS 95 + 97 P and of the future 15 searches are compared.
Neutralino annihilations should also produce an extra flux of positrons.
Those particles lose energy via synchroton emission and the inverse Comp-
ton scattering on both the CMB and diffuse starlight. As shown by E. Baltz
504 PIERRE SALATI
and J . Edsjo [49], primary fluxes are generally too small to be visible when
an isothermal halo model and canonical propagation parameters are as-
sumed. They are typically an order of magnitude or more smaller than the
Heat [50] measurements. Furthemore, the excess at 6-50 GeV in the Heat
data has not been confirmed by the recent measurements [51] that are now
in agreement with a pure secondary origin of the positron radiation.
F. Donato. N. Fornengo. P. SaiaU (1999)

103
10 2
10 1
"""'
I::
"-
>-(l) ----------------
d
..........
C'j
0.1
......
0
.......-
.......- 10-2
~
II:!
10- 3
..'
• ,\*}.
'.'
"
10- 4
p
10-5 ~~~~~~~~~~_LLllllm~~llW~WU~;·~:~:~~m_~~~
10- 8 10-7 10- 6 10-11 10-4 10-3 10- 2 0.1
~ji (Tji = 0.24 GeV) (m- 2 S-I sr- 1 GeV- I)
Figure 19. In this scatter plot, the antideuteron yield Nfj of Fig. 18 is featured against
the supersymmetric p flux (48). The antideuteron signal is estimated at solar maximum .
This corresponds to the AMS mission on board the space station. The p flux is derived
on the contrary at solar minimum, in the same conditions as the BESS 95 + 97 flights
whose combined measurements are indicated by the vertical shaded band for a p energy
of 0.24 GeV. The correlation between the antiproton and antideuteron signals is strong.
Travel at the centers of the Sun and of the Earth.

Neutralinos may be captured by celestial bodies as they cross them and
lose energy by scattering on a nucleus therein. The capture rate is the sum
of a pure geometrical
M
r geo ,...., 9.6 X 10 26 S-1 p~.3 V;OO ml~ M L:, (284)
o
and gravitational term
where p~.3 = m)(/0.3 GeV cm- 3 and V;oo is the neutralino velocity ex-
pressed in units of 300 km/s. The sum of the product (aiv)26 (Xi) / Ai over
the various nuclear species with averaged mass fraction (Xi) and atomic
number Ai is denoted by L:. Neutralinos concentrate at the centers of the
Sun and of the Earth where dense distributions built up. Those are strong
neutrino sources through the mutual annihilations taking place there. A
steady regime is soon reached where annihilations and captures balance
each other. In particular, free- escaping muon- neutrinos are emitted. A few
of them get transmuted into muons as they interact in the ground below
some terrestrial detector. This results into a flux of up-going muons that
may be observed through the Cerenkov light which the particles radiate as
they pass through water or ice.
The idea to instrument with optical modules a large domain in the polar cap
or in the ocean is strongly pushed forward. Those so-called neutrino tele-
scopes are actually mostly sensitive for large neutralino masses, a domain
out of the reach of the above- mentionned investigations. The neutralino
annihilation rate rain the solar core is just given by the trapping rate r c.
The subsequent neutrino flux v varies therefore as
(286)
The neutrino energy Ev scales typically as m x . The conversion of these neu-

trinos into muons proceeds through weak charged current exchange whose
cross section varies like G} mp Ev ex: m x . The flux of up-going muons does
not depend much therefore on the neutralino mass since both factors of mx
cancel each other. Finally, muons lose through ionization on the order of
0.26 GeV for each meter of ice or water which they cross. The effective size
of the conversion region increases with neutrino energy. The up-going muon
signal crudely scales as m)(. The larger the neutralino mass, the better the
limit on the interaction cross section of those particles. A word of caution;
506 PIERRE SALATI
the latter generally decreases with mx and neutrino absorption inside the
Sun tends to degrade the muon signal at high energy.
As shown in Ref. [52], the limit of,..., 2 - 4 X 103 muons km- 2 yr- 1 (EJ.L > 1
Ge V) set by the Baksan detector [53] just grazes the top of the set of super-
symmetric configurations in the (<p J.L ' m x ) plane. The few configurations
which are excluded by Baksan are also excluded by the direct searches.
Because the capture rate in the Earth is dominated by scalar interactions,
the correlation between the up-going muon flux from the Earth core and
the spin-independent cross section as! is strong. Such a correlation does
not exist in the case of the neutrinos originating from the Sun where the
neutralino capture also depends on axial interactions. The main source of
background arises from the showers produced by cosmic rays impinging on
the top of the atmosphere. In order to reject the downward-going muons,
several optical modules need to be hit during a single event, the time pa-
tern allowing then to reconstruct the trajectory. The next generation of
neutrino telescopes will be more sparsely instrumented. Muon ionization
implies therefore a larger threshold of ,..., 25 Ge V to be compared to the
present value of,..., 1 GeV . There is an irreducible background arising from
the cosmic ray induced shower neutrinos produced on the other side of the
Earth. With a 10 km 2 yr detector, supersymmetric configurations will be
probed at the 3-a level over a range of up-going muon fluxes extending
from ,..., 10 to 104 particles km - 2 yr -1.
Following a previous suggestion by A. Gould, a new population of neutralino
dark matter has been recently found - at least theoretically - in the solar
system [54]. Some neutralinos are gravitationally trapped by scattering on
the outer layers of the Sun and evolve on grazing and quite eccentric orbits.
Precession of the perihelion occurs as a result of the non-Coulomb nature
of the gravitational potential which the particles experience as they venture
below the solar surface. Perturbations due to planets, mostly Jupiter, can
make the orbits a little less eccentric so that the particles no longer intersect
the Sun. Protected against a new and fatal scattering, a new population
of neutralinos builts up, with very elongated orbits. It can persist in the
solar system for more than a Gyr. This new population intersects the Earth
orbit. This leads to an enhancement by at most a factor of,..., 2 in the rates
for direct detection. Because those solar system neutralinos have a lower
velocity than their halo partners, they are more efficiently trapped in the
Earth core - see Eq. (285). The capture rate and the up-going muon flux
sensitively depend on that velocity distribution. For some supersymmetric
configurations with mx = 60 - 100 GeV, the muon signal can exceed by
two orders of magnitude that predicted for halo neutralinos alone. The
net effect [55] of that new population is to shift upwards by at most an
order of magnitude the constellation of supersymmetric configurations in
the (<p Jl. , m x ) plane. They are now configurations exceeding the Baksan
limit. Those are nevertheless already excluded by the direct detection bound
on as/.
If the halo profile is cuspy, the collapse of the massive black hole that sits
at the galactic center generates the formation of a highly dense neutralino
spike. To illustrate how strong a neutrino source such a clump could be, let
us crudely estimate the corresponding up-going muon flux at the Earth. To
commence, the neutralino annihilation rate inside the central spike is just
set by the inverse of the age of the system, hence a value of rv 3 X 10- 18
s-l.
Problem nO 63 - Level [1] : Show that the total number of annihilations

taking place therein amounts to
Na rv 9 X 1043 S -1 \ -1
/\ m lOO , (287)
where A and mlOO have already been defined.
Let us assume in addition that neutralinos mostly annihilate into weak

gauge bosons. The latter subsequently decay into muon neutrinos
x + X -r W- + W+ -r /-l+ vJl. + ... (288)
with a branching ratio Bv rv 0.1.
Problem nO 64 - Level [1] : Infer a muon-neutrino flux at the Eartl1 of
<Pv = -BvNa
-2- = (-3
10 v s -1 cm -2) Am -1
lOO . (289)
47ff0
Two-body decay implies a typical neutrino energy of Ev rv Ew /2 rv m x /2.

The converted muon follows the impinging neutrino direction and keeps
typically half of its energy so that Ell m x /4. The muon attenuation
length because of ionization is
(290)
We crudely approximate the conversion cross section by
-38 2 ( Ev ) (291)
av-tJl. = 10 cm 1 GeV .
508 PIERRE SALATI
Problem nO 65 - Level [1] : Show that the neutrino to muon conversion

rate is given by the product
(292)
This leads to a muon signal of
(293)
to be compared to an atmospheric up-going muon background of rv 5 par-

ticles km- 2 yr- 1 above a threshold of 25 GeV. The neutrino signal from
the galactic center spike should be clearly visible by the next generation of
neutrino telescopes. A detailed analysis may be found in Ref. [41J. As Gon-
dolo and Silk conclude, haloes with inner cusps may have spikes so bright
that the absence of a neutrino signal already places upper limits on some
supersymmetric configurations, depending on the inner halo slope ,.
Acknowledgements: I would like to thank the organizers for their im-

pressive efficiency, their dedication and the kind and very warm atmosphere
which they succeeded to create during this Cargese School.
References
1. S. Weinberg, Gravitation and Cosmology (1972) Wiley, New York.
2. C.W. Misner, K. Thorne and J .A. Wheeler, Gravitation (1973) Freeman, San
Francisco.
3. L.D. Landau and E.M. Lifshitz, Classical Theory of Fields (1975) Pergamon
Press, Oxford.
4. R.M. Wald, General Relativity (1984) University of Chicago Press, Chicago.
5. S. Perlmutter et al. (TheSupernovaCosmologyProject), Astrophys. J. 517, 565
(1999) .
6. A. McKellar, PubIs. Dominion Astrophys. Observatory (Victoria, B.C.) 7, 251
(1941).
7. R.A. Alpher, H. Bethe and G. Gamow, Phys. Rev. 73, 803 (1948).
8. R.H. Dicke, P.J.E. Peebles, P.G. Roll and D.T. Wilkinson, Astrophys. J. 142, 414
(1965) .
9. A.A. Penzias and R.W. Wilson, Astrophys. J. 142, 419 (1965).
10. J.C. Mather et al., Astrophys. J. Letters 354, L 37 (1990).
11. G.F. Smoot et al., Astrophys. J. Letters 396, L 1 (1992).
12. Fixsen et al., Astrophys. J. 473, 576 (1996).
13. P. de Bernardis et at. (The Boomerang Collaboration), New Astronomy Reviews

43, 289 (1999).
14. N.A. Bahcall, J.P. Ostriker, S. Perlmutter and P.J. Steinhardt, Science 284, 1481
(1999).
15. R .B. Tully and J.R. Fisher, Astron. & Astrophys. 54, 661 (1977).
16. J.C. Tsai, Astrophys. J. Letters 413, L 59 (1993).
17. J.J. Mohr, B. Mathiesen and A.E. Evrard, Astrophys. J . 517, 627 (1999).
18. S. Burles and D . Tytler , Astrophys. J. 499, 699 (1998) and Astrophys. J. 507, 732
(1998).
19. P. Salati, P. Chardonnet, X. Luo, J. Silk and R. Taillet , Astron. & Astrophys. 313,
1 (1996) .
20. B. Paczynski, Astrophys. J. 304, 1 (1986).
21. A. Gould, Astrophys. J. 441, 77 (1995).
22. E. Aubourg, N. Palanque-Delabrouille, P. Salati, M. Spiro and R. Taillet, Astron.
& Astrophys. 347, 850 (1999).
23. P. Salati, R. Taillet, E. Aubourg, N. Palanque-Delabrouille and M. Spiro, Astron.
& Astrophys. 350, L57 (1999).
24. Ya.B. Zel'dovich, Adv. Astron. Astrophys. 3, 241 (1965);
H.Y. Chiu, Phys. Rev. Lett. 17, 712 (1966) ;
D.A. Dicus, E .W . Kolb and V.L. Teplitz, Phys. Rev. Lett. 39, 168 (1977);
P. Hut, Phys. Lett. B 69, 85 (1997);
B.W. Lee and S. Weinb erg , Phys. Rev. Lett. 39, 165 (1977).
25. A. Bouquet, P. Salati and J. Silk, Phys. Rev. D 40 , 3168 (1989).
26. K. Griest and J. Silk, Nature 343, 26 (1990).
27. A. Bottino , F .Donato, G. Mignola, S. Scopel, P. Belli and A. lncicchitti, Phys. Lett.
B 402 , 113 (1997).
28. R. Bernabei et at., Phys. Lett. B 389, 757 (1996).
29. P.F . Smith et at. , Phys. Lett . B 379, 299 (1996).
30. R. Bernabei et at. , Phys. Lett. B 436, 379 (1998).
31. A. Alessandrello et at., Phys. Lett. B 384, 316 (1996) .
32. L. Bergstrom, J. Edsjo and P. Ullio, Phys. Rev. D 58, 083507 (1998) .
33. L. Bergstrom, J. Edsjo, P. Gondolo and P. Ullio, Phys. Rev. D 59, 043506 (1999).
34. P Ullio, astro-ph/9904086.
35. M. Davis, G. Efstathiou , C.S. Frenk, and S.D.M. White, Astrophys. J. 292 , 371
(1985).
36. E.A. Baltz, C. Briot, P. Salati, R. Taillet and J. Silk, Phys. Rev. D 61, 023514
(2000).
37. J .F. Navarro, C.S. Frenk and S.D.M . White, Astrophys. J. 462 , 563 (1996) .
38. A.V. Kravtsov, A.A. Klypin and A.M . Khokhlov, Astrophys. J. Supp!. 111, 73
(1997) .
39. Moore et at., Astrophys . J. Letters 499 , L 5 (1998).
40. A.M. Ghez, B.L. Klein, M. Morris and E.E. Becklin, Astrophys. J. 509, 678 (1998).
41. P. Gondolo and J. Silk, Phys. Rev. Lett. 83, 1719 (1999) .
42. L. Bergstrom, P. Ullio and J. Buckley, Astropart . Phys. 9, 137 (1998).
43. H. Matsunaga, Ph.D. thesis, University of Tokyo, 1997.
44. S. Ahlen et at., Nucl. lnstrum. Meth. A350, 351 (1994) .
45 . A. Bottino, F. Donato , N. Fornengo and P. Salati , Phys. Rev. D 58, 123503 (1998).
46. L. Bergstrom, J. Edsjo and P. Ullio, astro-ph/9902012 (1999).
47. J.W. Bieber et al., Phys. Rev. Lett. 83 , 674 (1999).
48. F. Donato, N. Fornengo and P. Salati, preprint lapth-722/99 and FTUV /99-9 and
astro-ph/9904481 , to be published in Physical Review D.
49. E.A. Baltz and J. Edsjo, Phys. Rev. D 59, 023511 (1999).
50. S.W. Barwick et al. (Heat Collaboration), Astrophys. J . Letters 482 , L 191 (1997) .
51. M. Boezio et al. (Caprice Collaboration), Proceedings of the 26th lCRC, contribu-
510 PIERRE SALATI
tion OG.1.1.16 (1999), Salt Lake City, USA.

52. L. Bergstrom, J. Edsjo and P. Gondolo, Phys. Rev. D 58, 103519 (1998).
53. M.M. Boliev et al. (Baksan Collaboration), Nucl . Phys. B (Proc. Suppl.) 48, 83
(1996), TAUP 95, Toledo, Spain, 17-21 September 1996, ed. A. Morales.
54. T . Damour and L.M. Krauss, Phys. Rev. Lett. 81, 5726 (1998)& Phys. Rev. D 59,
063509 (1999).
55. L. Bergstrom, T. Damour, J. Edsjo, L.M. Krauss and P. Ullio, Journal of High
Energy Physics 010, 9908 (1999).
INDEX
Abelian gauge invariance, 95-100, Couplings (cont'd)

343-344, 367, 381 quartic, 55-56, 66, 72
Antimatter, 397-401, 407, 500-508 anomalous, 55-63, 72
CP-violation, 171-201, 221-222, 232,
398-402, 406-414
Baryogenesis, 397- 414
Affleck-Dine, 413 basic formalism, 172-177
electroweak, 403-410, 414 in B-meson system, 190-201
GUT,41O-412 in decay amplitude, 179
Baryon asymmetry, 399- 403, in neutral Kaon system, 180-190
408-410, 413-414 in oscillation, 179
Bekenstein entropy, 387 in rare decay, 188- 190
B-hadron lifetime, 220-221 in standard model, 186-188,
Black holes, 386- 390, 494 190- 198
Bogomolnyi bound, 346-349
Borel transformation, 252- 254, 259 Dark matter, 486-488, 492-499
BPS-states, 351-354, 357, 383 baryonic, 469-475
cold, 476, 492
CDF-detector, 214- 215 hot, 101
Chandrasekhar limit, 433-434 supersymmetric, 500-504
Charge asymmetry, 30, 33- 34, 37 D-branes, 383-393
Chargino, 87 Deep inelastic scattering, see HERA
Chiral symmetry, 329-337 D0-detector, 215-216
Chooz experiment, 139 Duality
CHORUS experiment, 146-151, 155 connections, 378-381
CKM-matrix, 171, 187- 198, 220- 221, electric-magnetic, 340-343,366,381
273-274 in field theory, 369- 371
CNGS project, 161- 162 in Yang-Mills theory, 353- 355
COBE, 437, 441 Montonen-Olive, 354- 355
Cosmological constant, 281 , 422-429, transformation, 365
431, 436, 461 Dyon, 346- 349, 364
Couplings
trilinear, 23-26, 55, 66, 72-73 Einstein ring, 468, 471-473
511
512
Electron-positron annihilation ICARUS project, 163-164

event shape, 261-265 Instanton, 251, 255
into fermion pair, 3-7, 30-35, 71,
75 Jet production, 216-218
into hadrons, 6-7, 247-249
into W+W-, 12-14, 20, 39, 74 KARMEN experiment, 140-145
into ZO ZO, 27-29 K2K project, 156-159
Electron-posi tron linear collider,
65-91 Lensing
Electron-proton scattering, gravitational, 461-469
see HERA micro, 471-473
Extra dimensions, 319-323 LEP accelerator, 2-3
Leptoquark, 209, 227, 410-412
Friedmann-Lemaitre model, 418 Lineshape, 30-32
LSND experiment, 140-145, 166
GALLEX experiment, 114-115
Gamma ray, 495-500 MACRO experiment, 135
Gaugino, 89-90, 488 Maldacena conjecture, 392
Gluon density, 207 Microwave background, 399-400, 425,
Grand unification, 67, 75-77, 85, 436-452, 503
89, 301-303 Minimal supersymmetric standard
supersymmetric, 300-306 model (MSSM), 285-297
Gravitino, 314-319, 413 flavor physics, 289-290
precision tests, 290- 291
Hawking radiation, 386-389 spectrum, 292-297
HERA MINOS project, 159- 161
CC cross-section, 207-208 Monodromy, 362-367
experimental results, 203-211 Monopole
NC cross-section, 206-208 condensation, 367-369
Higgs boson, 77-78 magnetic, 341-343, 368-369
coupling, 81-82 Yang-Mills, 343-347
decay, 48-50, 78, 81
mass limit, 46-48, 66, 229-231, Neutralino, 87
297, 409 annihilation, 495, 502-503
search, 45-48, 66, 229-231, 297 capture, 505-507
strahlung, 47-48, 77-79 galactic, 491-494
supersymmetric, 83-85 mass, 501- 502, 505
width, 79-80 search, 488-492
Higgsino, 488 synthesis, 475-487
Homestake experiment, 113-114 Neutrino
Hubble constant, 101-102, 435-436, atmospheric, 126-138, 481
460- 461, 469 angular distribution, 128-129
513
Neutrino (cont'd) Renormalon, 249, 255- 260

flux, 127- 128 R-parity, 229, 487
origin, 126- 127 R-symmetry, 358, 364, 392
cosmic , 481- 482 Running coupling constant
Dirac, 101, 487 QCD, 246- 248, 390
heavy, 481, 485- 487 QED, 8- 10
helicity, 101 supersymmetry, 360-361
Maiorana, 101, 487 Rydberg energy, 441
mass, 101- 103, 165-166
oscillations, 101- 167, 412 Sachs-Wolfe effect, 442- 443
accelerator searches, 140- 155 SAGE experiment, 114- 115
experiments, 104- 106 Slepton, 87-90
in matter, 106- 109, 122- 123 Squark, 87-89, 491
in vacuum, 103- 104, 119- 121 Standard model, 267-275
scattering, 478- 488 naturalness, 279-280
solar, 109- 126 precision measurements, 30- 45
data, 117- 119 problems, 275- 279
experiments, 113- 117, 123-126 Standard solar model, 109- 113, 116
String, 371- 375, 393
NOMAD experiment, 146- 147, unification, 319- 323
151- 155 superstring, 376- 378
NUMI project, 159- 160 Structure functions, 203- 206
Super-Kamiokande, 116- 119, 123,
OPERA project, 164- 165 130- 131, 135, 138, 156- 157,413
Supernovae search, 429-436
Photino, 488 Superstring, 376-378
Point splitting, 95-100 Supersymmetric particles, 67- 68,
Polarization asymmetry, 30, 35 87-91
search, 298-300, 487-508
Quantum chromo dynamics (QCD) Supersymmetry, 67, 85- 91, 229,
Lagrangian, 245 349-352, 392
A QCD,246 algebra, 282-283
nonperturbative hadronization, breaking, 67-68, 86, 229, 287-288,
245- 265 310- 313
running coupling constant , motivation, 281
246-248 naturalness, 284- 285
Quark confinement, 368
Quintessence, 423- 424, 436 T-duality, 375- 376
Technicolor, 227- 228
Robertson-Walker TEVATRON
coordinates, 423, 429 experimental results, 216-231,
metric, 418- 420 243-244
514
TEVATRON (cont'd)
jet production, 216-218
main injector, 214
O-angle, 347-349, 361
Thermal quenching, 478-481
Thrust, 261-265
t-quark, 66-68
decay, 69
dipole moment, 70
magnetic moment, 70
mass, 70- 71, 222, 225- 226
Vector boson production, 235- 244
W-boson
decay
hadronic, 17-21
leptonic, 14-15
semi-Ieptonic, 15-17
mass, 28-30, 37-45, 73, 208,
224-225
production, 218-219, 223
width, 21-23
Z-boson
limits on extra -, 11-12
production, 218- 219, 223
Zino,488

Particle Physics: Ideas and Recent Developments

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Particle Physics: Ideas and Recent Developments

Uploaded by

Copyright:

Available Formats

Particle Physics: Ideas and Recent Developments

NATO Science Series

A. Life Sciences lOS Press

1. Disarmament Technologies Kluwer Academic Publishers

Series C: Mathematical and Physical Sciences - Vol. 555

Springer-Science+Business Media, BV.

ISBN 978-0-7923-6436-8 ISBN 978-94-011-4128-4 (eBook)

Printed an acid-free paper

AII Rights Reserved

Preface ................... . . . ............................... . ........ VB

Results from LEP 200 ............................... 0 . 000... 0 0. ...

Baryogenesis ........................................................ 397

Cosmology and Dark Matter ......................................... 417

Pierre Binetruy Daniel Froidevaux

Raymond Gastmans Jean-Marc Gerard

Jean Iliopoulos Maurice Levy

Lecture 1 is devoted to some of the topics that can be studied from

2. The LEP accelerator

in 240 MW of electric "plug" power. Furthermore the inefficiency of normal

Period Energy (Ge V) Luminosity (pb -1)

1995 130/136 6.2

TABLE 1. Energies and integrated luminosities

3. Fermion pair-production at LEP-2

Figure 1. First order Feynman diagrams for e+e- ~ If.

The total annihilation cross-section (see ref. [1]) e+e- -+ f f has a

Figure 2. e + e - annihilation cross-section as a function of the collision energy, for the

this energy the cross-section is also dominated by fermion-pair production

··········•······ ........... Lqq (ISR)

120 140 160 240

Figure 3. The e+ e- annihilation cross-section at LEP-2 energies

(2) s'/ s > 0.01 (inclusive)

OPAL 183 GeV

50 100 150 200 100 150 200

Figure 4. s-prime distribution for e+e- -+ If from OPAL [13] .

(J (e +e -' >ff) LEP (preliminary) A (e+e'->ff) LEP (preliminary)

• e+e ....> ~l+-

which are covered below.

3.1. MEASUREMENT OF THE I'-Z INTERFERENCE

3.2 . MEASUREMENT OF THE RUNNING OF (XEM(S)

For non-radiative leptonic events the dominant diagram is 'Y exchange in

...1 OPAL <> .......... LEP 1

91.17 91.18 91.19 91.2 91.21 91.22

100 0 20 40 60 80 100 120 140 160 180200

Figure 9. OPAL measurements of (tern together with measurements at lower energies

Two analysis have been done:

a;~(157.42 GeV) = 119.2:!=g

a;~(157.42 GeV) = 119.9::~ :~ ± 0.1

3.3. LIMITS ON FOUR-FERMION CONTACT INTERACTIONS

New interactions at a much higher energy-scale A, can show up at lower

3.4. LIMITS ON EXTRA Z BOSONS

Extra Z bosons are present in many GUT theories, such as E6 (several

M~(sequential) > 898 GeV (at 95%)

4. Measurement of the WW production cross-section

Figure 11. The CC03 diagrams for WW production in e + e - interactions.

Each W then decays to two fermions, leading to a 4-fermion final state.

Figure 12. Diagrams contributing to the final state e+e- -t udl.LlJI-"

N~§03 = number of events attributed to 4f, non-CC03 diagrams,

The "4f-CC03 correction", N~fo3 / L, amounts to about 1% at most. The

4.1. FULLY LEPTONIC W DECAYS

These events are characterized by

30 90 120 150 l80

>Q) (b) • Data Iv lv L3

Figure 14. Lepton energy in purely ieptonic decays of the Wi s.