Professional Documents
Culture Documents
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
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...
A C.I.P. Catalogue record for this book is available from the Library of Congress.
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
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
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
f
"Z
e f
<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)
-:D
~
b 10' ••••• a(e+e- ~ X) (pb)
ww,~~~~------J
-,
10
10
.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
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].
RESULTS FROM LEP 200 7
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.
15 f
20
III
10
-10
-IS
.
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]) .
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.
• -LEPl+lJO..l72GeV
0.75
0.5
-0.5
-0.75
·1
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
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
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.
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
RESULTS FROM LEP 200 11
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.
where the W's are "real" (resonant). There are 3 main diagrams (called
CC03, Fig. 11) which contribute to this channel:
e'
N obs - N baek - N 41
eeD3
O'ee03 = L
fee03
RESULTS FROM LEP 200 13
'K~ e
"
W
G
e " 3
2 2
graph 4 graph 5
2 2
graph B graph 9
2
graph 10
N 4/ = L[ E4/
cd13 MC MC]
- Ecc03(} cc03
where
Nabs = number of observed events,
Nback = number of events from non-like WW diagrams
1
:! ALEPH
'"
t,o' o
1
Me tWW-tlvlV + BIN:kgmllmJs,
MC(WW-+N"',
'0 2
Me {Blld:gmwuls}
/JArA
'0
.,
'0
Figure 13. Acoplanarity angle of the two charged leptons in purely leptonic decays of
the WiS .
p reliminary
<Il
C50
Q)
>
w
'0
~cut
0525
.0
E
•
:::l
Z
20 40 60 80 100
Elepton [GeVJ
Figure 15. An ALEPH WW event where the W- decays leptonically and the W+
decays hadronically.
20
c 11 >,t. -' ALEPH
~ 18 .Js= -16 1 CF;:V
• 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 .
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).
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
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.
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.
RESULTS FROM LEP 200 21
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
....
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
10 11 12 66 68 70
Figure 21. The W ieptonic and hadronic branching ratios for the four LEP experiments
together [15].
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
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
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
RESULTS FROM LEP 200 23
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
W+(q)
V(p) ~ = --- iegvwwr v:6/>(q,q,p)
"\. g-yww = 1 .
W-(q) . 9zww = cot Ow
6z = gzww - cot Ow
e----~------~~~----+----
f
Figure 23. The angular variables used in the analysis of the triple gauge couplings.
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 ~
1
150
>
OJ
'0 100
.ll8
SO i. 50
.{l,S 0
<>'1 O. O.S 2 4
CQ~e <I>
Figure 24. The distributions of the angular variables for the analysis of the trilinear
couplings obtained by the L3 collaboration.
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
\
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.
RESULTS FROM LEP 200 27
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.
~
'.
,
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
(jZZ (pb)
l3 O 75 .;. 1)· 15
· - r' 11
common 0.015 pb
·lldof = 0.63 / 3
SM 0.65 ± 0.01
I ••• • I .---t- • • I •• , , I
o. 1.
all (pb)
LEP - preliminary
I.S , - - - - - - - - . ! . - - - - - - ' ' - - - - - - - - - - ,
t' Data
_ Standard Model
0 .5
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
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:
_ 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
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"
(7
o 8r~
ff (8 _ M~)2 + (~)2
12rr rer j
M~G
#(1 - 4sin Otff(M~))
2
#
RESULTS FROM LEP 200 33
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
Ai _ O"F - O"B
FB - O"F +
O"B
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
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
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
10.0
10
J:
g r:n.. .. 1i4 :t 5C;I!'V
E 10
lOO
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
RESULTS FROM LEP 200 35
PT(M~) -AT
3
A~POL(M~) --A
4 e'
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
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].
RESULTS FROM LEP 200 37
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.
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
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
RESULTS FROM LEP 200 39
~ 3
~
man
Fi All] LEP 161 GcV W
LEP EW WOf'klnl Group
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
~ 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].
± 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
RESULTS FROM LEP 200 43
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 .
The error on the beam energy enters directly into the the error of the
W mass, through the kinematic fits:
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.
87l'v 2
m2
H
<
- 3l A2
v:::::: 246GeVjc2 .
og-;r
If we take A up to the Plank scale the limits are
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
RESULTS FROM LEP 200 47
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.
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.
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%).
Z -t lIii
H -t bb
H -t bb
or
Z -t qij
H -t T+T-
,
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.
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
RESULTS FROM LEP 200 51
expected.
500 observed *
400 ALEPH
~
V)
300
DELPHI
o
s::: 200 L3
E OPAL
:J 100
....J
50
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
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
4. ALEPH Collaboration, Phys. Lett. B399 (1997) 329
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
9. L3 Collaboration, Phys. Lett. B407 (1997) 361
10. L3 Collaboration, Phys. Lett. B433 (1998) 163
11. OPAL Collaboration, Phys. Lett. B391 (1997) 221
12. OPAL Collaboration, Eur. Phys. J. C2 (1998) 441
13. OPAL Collaboration, Eur. Phys. J. C6 (1991) 1
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.
16. ALEPH Collaboration, Phys. Lett. B401 (1997) 347
17. ALEPH Collaboration, Phys. Lett. B415 (1997) 435
18. ALEPH Collaboration, Phys. Lett. B453 (1999) 107
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
24. OPAL Collaboration, Eur. Phys. J. C1 (1998) 395
25. OPAL Collaboration, Eur. Phys. J. C8 (1999) 191
26. ALEPH Collaboration, Phys. Lett. B422 (1998) 369
27. ALEPH Collaboration, CERN-EP-98-178
RESULTS FROM LEP 200 53
W. JAMES STIRLING
Departments of Physics and Mathematical Sciences
University of Durham
Durham DHl 3LE, u.K.
AND
ANJA WERTHENBACH
Department of Physics
University of Durham
Durham DHl 3LE, U.K.
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
J.-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 55-64.
© 2000 Kluwer Academic Publishers.
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.
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
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)
58 W. JAMES STIRLING AND ANJA WERTHENBACH
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 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.
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.
60 W. JAMES STIRLING AND ANJA WERTHENBACH
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.
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].
"
•••• ~-'
-, ..- / .....
.' •• 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
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.
62 W. JAMES STIRLING AND ANJA WERTHENBACH
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
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.
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'.
64 W. JAMES STIRLING AND ANJA WERTHENBACH
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
1. Synopsis
1.-1. Aubert et a/. (eds.), Particle Physics: Ideas and Recent Developments, 65-93.
© 2000 Kluwer Academic Publishers.
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.
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.
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.
PHYSICS WITH AN e+e- LINEAR COLLIDER 69
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.
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.
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:
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]
PHYSICS WITH AN e+e- LINEAR COLLIDER 73
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,
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
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
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}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]
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
PHYSICS WITH AN e+e- LINEAR COLLIDER 81
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:
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.
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
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,
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:
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.
match surprisingly well, the difference being less than about 2 per-mille.
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 ~~--~~~~~~~~
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].
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
PHYSICS WITH AN e+e- LINEAR COLLIDER 89
mixing, these particles may be pair produced even at a 500 GeV collider:
i,j = 1,2
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.
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:
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.
PHYSICS WITH AN e+e- LINEAR COLLIDER 93
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
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
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
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)
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. ,
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
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
J.-J. Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments, 101-169.
© 2000 Kluwer Academic Publishers.
102 L.DILELLA
(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)
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
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)
Pa /3 = ~ sin2 (28) .
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).
v v v e-
~
va
~
,
v.
•
~
e,N
:z
I
~
:w
I
>-~-<
e" e-
e,N e- ve
a) b) c)
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
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
where
e= 2VwE ~ 1.526 x 1O- (ZjA)pE eV2
7
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
NEUTRINO OSCILLATIONS 109
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.
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
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:
NEUTRINO OSCILLATIONS 111
(i) The pp cycle, responsible for 98.5% of the Sun luminosity. This cycle
involves the following reactions:
p+p -+ e+ + Ve + d (15)
(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
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.
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
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.
(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) = 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
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
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)
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).
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 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)
which implies that Be7 must exist in the Sun. This in turn implies the
occurence of the reaction
(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.
(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
-
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].
NEUTRlNO OSCILLATIONS 121
~ 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
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
NEUTRINO OSCILLATIONS 123
Rd-Rn
Rd + Rn = -0.065 ± 0.031 (stat.) ± 0.013 (syst.) ,
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
NEUTRINO OSCILLATIONS 125
10000
iii
~
1000
>co
-
~
N
0
ci
100
J!l
c:
::J
0
t.l
10
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
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
NEUTRINO OSCILLATIONS 127
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%.
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
NEUTRINO OSCILLATIONS 129
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
.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
Cose Cose
Il: 1.5 , - - - - - - - " I l : 1.5 , - - - - - - - - - - - - ,
(e) Kllllioklulde ~ Super Kllmioklulde
M1lti.QoV
Cose Cose
e-like
0.&
-0.5
5'
+
~ -~O 'l 10
9 1~-r-r~~~-'~~~",---------,
2. IJ.-like
0.5
FC PC
+' ..•...~' ""., ..
,w•••••••· m•••••m.wmN
10 . ,
10
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
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.
NEUTRINO OSCILLATIONS 133
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.
-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).
NEUTRlNO OSCILLATIONS 135
....
-. If) ....,............... - .! ........... :.. : ..... , ...... \ ........... .
"" <;I . •
~ ,. ~~
i:: .v ; ~
:...
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 •
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
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
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 .
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].
has not yet reached the sensitivity to oscillations of the Chooz experiment.
140 L.DILELLA
1 r--..-------~~~~------------,
~ 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].
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
NEUTRINO OSCILLATIONS 141
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.
LSND KARMEN
Angle between
v direction and
proton beam
Data taking period 1993-98 Feb. 97-Feb. 99
Protons on target 2.9 X 10 22
NEUTRlNO OSCILLATIONS 143
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.)
§
~. '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.
NEUTRINO OSCILLATIONS 145
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.
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
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
CHORUS
.,001 baI:
i DUOIl 8p«tramel:er
Vlewaf
microsc~p.
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.
! ~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:
The ratio < U T > / < 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
~
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
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
NEUTRINO OSCILLATIONS 153
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'
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,..
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
NEUTRlNO OSCILLATIONS 155
90% confidence level upper limit using the method of Ref. [56] is [60]
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.
NEUTRlNO OSCILLATIONS 157
.'
~.('...... 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
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.
NEUTRINO OSCILLATIONS 159
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
~
"-
~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.
WLS fi bu readout
Magnetized Fe Plates
4S6 Layers x 2.54 cm Fe
5.4 kT Total Mass
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
NEUTRINO OSCILLATIONS 161
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 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)
.2 (20) (tl.m)
NT = 1.61 A sm 222/
L ¢v,. (E)uT(E) dE
E2 . (46)
dm 2 (eV2) Nr
10- 3 2.48
3 X 10- 3 21.7
5 x 10- 3 58.5
- \"I.
J mm 3 IT""
1
so 00 50 <1-1111)
3 X 10- 3 4.5
- 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.
NEUTRlNO OSCILLATIONS 167
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
T.NAKADA
CERN EP-Division
CH-1211 Geneva 23, Switzerland
and
Institute of High Energy Physics
University of Lausanne
CH-l015 Dorigny, Switzerland
1. Introduction
1.-1. Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments, 171-202.
© 2000 Kluwer Academic Publishers.
172 T . NAKADA
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
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
hence
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
can be written as
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
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
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
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.
d2 a(t) da(t)
-;Ji2 + 2i A ~ + A12A21 - A
( 2) a(t) = 0 (7)
where C± are arbitrary constants which can only be defined by the initial
condition. For b( t) , we obtain
176 T.NAKADA
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)
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) .
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
and
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.
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:
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.
PHYSICS OF CP VIOLATION AND RARE DECAYS 179
~ ((~;) =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
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)
and
TL =rL1 = (5.17 ± 0.04) x lO-8 s
PHYSICS OF CP VIOLATION AND RARE DECAYS 181
i.e.
tlF = Fs - FL = (1.1174 ± 0.0010) x lO lD s-l
.,.
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.
and
arg € = (43.50 ± 0.08t.
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
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·
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.
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
PHYSICS OF CP VIOLATION AND RARE DECAYS 185
is totally dominated by the elastic scattering at the energy scale of the kon
mass. Similarly for the nOn o final state, we have
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+_
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
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:
so that
ICPr+ 2cpo + -OT - Ocpl < 0(10- 5 ) .
Thus ICPr + 2cpo + OT - Ocpl « lEI, justifying the approximations made
before.
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
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
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 .
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
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
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.
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.
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
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•
PHYSICS OF CP VIOLATION AND RARE DECAYS 193
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
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
and
194 T.NAKADA
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 .
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)
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
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
PHYSICS OF CP VIOLATION AND RARE DECAYS 197
~:../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
where
LC: _ = ~ x A(BO -t D*+7r-)
D 7r+ ( A(BO -t D*+7r-)
and the phase of Lg:- 7r + is given by
From the two time dependent decay rates, we can extract 4>3 - 24>1.
Note that
and
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
PHYSICS OF CP VIOLATION AND RARE DECAYS 199
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
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.J. CASHMORE
CERN
CH-1211 Geneva, Switzerland
1. Introduction
203
J.-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 203-212.
© 2000 Kluwer Academic Publishers.
204 R.J. CASHMORE
-l = sxy (2)
P.q
Pk (3)
k
q:
:,,(, w, z
I
(4)
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.
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
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
.-0.00253
A:8 ~=_.-~ ~ ~ ! - .~--
.·0.0032
1.:8 [ ............. ------ ....--------
6~8 ~:~:-==
o:g
_______.
o.g [ - - _a-.........____...... -'_L..~---'-
_.~
__.....-. ______ ..... ____ ._.....__
~...-
J _ ,a_AJ. L L L ___ ~ __
.-0.001
.·0.013
oi ~_..----.___.--.-.--------......----..-------~.-+-- x-0.032
Aj ~ _______.-....__ •__... ,. _______ L..a_..e__*_. . . . _ - - - - - - - . . _ . . . _ _ _ _ x.O.05
- -- _._
+ .. . •
A:gC.
--
.-0.'5 J
6:g [" ~.--.---------••--••--- - ---- ----.-----.-------------··------:-.0.55 J
... ..........----------- .
1d 1do 1dooo 1doooo
Ci (GeV')
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)
HERA - EXPERIMENTAL RESULTS 207
,,-... 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.
~ 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.
HERA - EXPERIMENTAL RESULTS 209
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
Figure 6. Measured NC cross section compared to the SM prediction for three assump-
tions of the ZO masses.
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.
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.
HERA - EXPERIMENTAL RESULTS 211
"
~
... 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)
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
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.
© 2000 Kluwer Academic Publishers.
214 H.E. MONTGOMERY
.
"ONIH f SOt,n .
'
.~.: ..
"J'
;".
'. -;,"J .
.,~:
• '- d
P- ....,l......,_
( ..... ) 0
','
~:.:
':" W-Jl~:::::::"'.,:t:F~~~~~~ ,.
(~)
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
"*"JETRAO
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.
-4
10
-5
10
·6
10
lO .L---L_..L....--J.._.L---J.._.L---'-_~--'--------'
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.
CDF Preliminary
0.055
.. ---- ...... .._.,._..... -.... .... . .
·~ -~ ·· o---a ·····
,,_-... .. ____ ....................... -4--,; .... -a'...... ----- ... ---.. -.~- ......-.............. _......-~ ..
0.050 •
0.045
. ---~---- ... -~..--------.- .. - -_... _---_._----------_..
Jet E~1n = 30 GeV
0.040
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.
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
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
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 221
CDF B Lifetimes
I I
0.5 1 1.5
--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)
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
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 223
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
:>
(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
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 227
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.
'"~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.
SUGRA I
-± 70 210
Xl
9 270 390
tl(-+ bX~) 170
GMSB
150 265
120
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 229
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.
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.
5. Conclusions
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
RECENT RESULTS AND PERSPECTIVES AT CDF AND D0 233
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
1999, World Scientific, 1999.
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
1999, World Scientific, 1999.
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
Department of Physics
University of Durham
Durham DHl 3LE, U.K.
AND
W.J. STIRLING
Departments of Physics and Mathematical Sciences
University of Durham
Durham DHl 3LE, U.K.
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
J.-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 235-244.
© 2000 Kluwer Academic Publishers.
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)
(.) (b)
Figure 1. a: Real diagrams for the LO process (2), b: One of the diagrams presenting real
gluon emission contributing to DLLA.
(4)
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)
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.
238 A. KULESZA AND W.J. STIRLING
:;--
G
"0
'-
b
10 5
"0
2- 10 4 _ DlLA
;
:::- 10 3
_ ___ O{a.}
10 2
10
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
(7)
HEAVY GAUGE BOSON PRODUCTION ... 239
(8)
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
-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
240 A. KULESZA AND W.J. STIRLING
_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
HEAVY GAUGE BOSON PRODUCTION ... 241
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
242 A. KULESZA AND W.J. STIRLING
"fl
00 (_2>.)(N-l) N-l (
(N - I)! ~O
N - 1)
m
[
L N - 1 - m 2TN+m + LTN+m-l
]
. (13)
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)
.,.
"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.
5. Conclusions
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,
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
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.
© 2000 Kluwer Academic Publishers.
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)
2
A QCD
1) ~ (0.2 GeV)
= Q 2 exp ( - f300!s(Q2) 2
. (1.3)
1,Z
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)
I{1.2
MAliK I
+- MAliK IILCIII
• I MilA
jI;.m (a.V)
Figure 2. Dependence of Re+ e- (s) on the center-of-mass energy. Solid lines correspond
to perturbative QeD fits.
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
(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.
(3.1)
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
(3.3)
(3.4)
THE THEORY OF STRONG INTERACTIONS 251
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
(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
N!
E
PT ~-~=-""'--'----------
best approx,
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
(3.1O)
(3.11)
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)
(3 .13)
(3.14)
and integrate over fluctuations a(x}. Taking into account that 6ASYM[A] =
o forA = A* we obtain
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)
(qJ-lqv - gJ-lvq2)II(l} = -i J
d4 xe- iqX (0IT {JJ-I(x)Jv(O)} 10). (4.2)
256 G.P. KORCHEMSKY
= 1 + L a~+1(Q2)
00
+ 2
(4.5)
(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
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
(4.9)
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)
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)
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)
Instanton-antiinstanton
singularities at cr=4lt. 81t ...
UV renormalons at IR renormalons at
cr=-n/~. n;1,2, ..... cr= n /~. n= 2 , 3 , " .
(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
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
(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
THE THEORY OF STRONG INTERACTIONS 263
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)
(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
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?
THE THEORY OF STRONG INTERACTIONS 265
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.
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
[, y M = _!GJJvAG A
4 JJV
-!4 WJJvIWIJJV _ !BJJV B
4 JJV , (2)
where
(3)
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
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.
(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).
The last but not the least important ingredient of the 8M is a complex
spin-O SU(2)-doublet, the so-called Higgs field:
which appears not only in the part of the Lagrangian containing the spin-O
fields gauge interactions and self-interactions, depicted in Fig. 3:
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.
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 271
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.
(11)
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].
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.
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 273
Stanford 1999
Measurement Pull
;':'"
~:..."
-3 -2 -1 0 1 2 3
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)
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 275
As we shall see, all these constraints can be violated if we extend the SM!
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,
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 277
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
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 279
600
:I: 400
:::a
200
(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.
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 281
C- 1
A cosm '"< Mp
F
rv
10- 4 eV. (28)
(29)
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 283
- - - - 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
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:
(36)
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.
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)
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
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.
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 289
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.
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)
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)
(55)
where
m~ = -m~ (tan{3 + ta~(3) . (58)
where the mixing angle a is conventionally chosen such that - ~ ::; a ::; 0
and is given by
(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
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.
SUPERSYMMETRlC EXTENSIONS OF THE STANDARD MODEL 295
(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
( M2 V2mwsin,8 ) (74)
V2mw cos,8 J1. '
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
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
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 297
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
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 299
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
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
SUPERSYMMETRlC EXTENSIONS OF THE STANDARD MODEL 301
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.
(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
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)
(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]:
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.
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
(88)
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
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
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 307
(93)
dMA bA 2
~ = 81r29AMA, (A = 1,2,3), (94)
and they are immediately solved with the boundary conditions (90), to give
(95)
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
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
SUPERSYMMETRIC EXTENSIONS OF THE STANDARD MODEL 309
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)
Mt
1 < tan/3 < Mb . (103)
The proof relies on the relation, derived from the minimization of Vo( Q):
(104)
310 FABIO ZWIRNER
(106)
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)
(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.
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)
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
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.
8. Concluding remarks
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.
SUPERSYMMETRlC EXTENSIONS OF THE STANDARD MODEL 325
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.
SUPERSYMMETRlC EXTENSIONS OF THE STANDARD MODEL 327
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
F. SANNINO
Yale University
Department of Physics
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
J.-J. Aubert et al. (eds.), Particle Physics: Ideas and Recent Developments, 329-337.
© 2000 Kluwer Academic Publishers.
330 F. SANNINO
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
(5)
(6)
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,
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
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
ENHANCED GLOBAL SYMMETRIES 337
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
J.-J. Aubert et al. (eds.), Particle PhYSics: Ideas and Recent Developments, 339-395 .
© 2000 Kluwer Academic Publishers.
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].
(2.4)
(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)
(2.12)
47r r3
342 ELIAS KIRITSIS
eg = 211"nli (2.16)
(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)
(2.19)
SUPERSYMMETRY AND DUALITY 343
'Y I 0 I 1 0
w± I ea I 1 ±1
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
(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.
[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.
(4.1)
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
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)
(4.4)
(4.5)
Oe 471")
(q,g)= ( ne+ 271"m, 7 m (4.6)
(4.8)
M 2: ae In + mrl (4.9)
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
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)
A a:m =.J2
1 [Qlm
a: + €a:/3 Q2m]
/3 '
Bm
a:
1 [Qlm
=.J2 a: - €a:/3
Q2m]
/3'
(5 7)
.
{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)
(5.11)
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
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
( aC db) , ad - bc = 1 , a, b, c, d E Z (6.3)
(6.4)
Exercise: Show that the BPS mass formula is invariant under the
SL(2,Z) duality.
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
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 do this analysis for general N although eventually we will be interested in
N=2.
SUPERSYMMETRY AND DUALITY 359
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
Exercise: Find the unbroken global symmetry for G=SU(3), SU( 4).
eff
g
m E
1 1
g2(J.L) = 95 - 1
871'2 ~ bi log
(J.L 2 + m?)
A2 I
(7.7)
1 1 2 IAI2
-2-
geff
= 2"
go
+ -8
7r
2 log A2 (7.9)
1 1 IAI2
-2- = 4 2 log A2 (7.10)
geff 7r
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
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
SUPERSYMMETRY AND DUALITY 363
Q
8
(7.17)
(7.18)
i A2 + A
Y(A) = 27r log A2
2 00
~ Ck
(A)4k
A (7.19)
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.
(7.20)
S = _1_Im
3~
I r(a)(F + iF)2 = _1_Im
~~
I r(a)(2F2 + 2iF F) (7.21 )
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.
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)
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)
(7.28)
vf2
A(u) = --;-
/1
-1
";x -
dx ~ =
V2(1 + u) F (-2'12'11; 1 +2)
U
u (7.31)
In - du -
v 2M M + m dAD = 0 , AD M = AD M = 0 (8.2)
(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
SUPERSYMMETRY AND DUALITY 371
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
(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.
--.~ 0
-'- ---------,..,
-
RI 0-
-r R-O
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)
(lOA)
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
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].
+ ......
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)
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.
6These are scalars that come from the tenth components of the gauge fields in ten
dimensions. These expectation values are called Wilson lines.
SUPERSYMMETRY AND DUALITY 379
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
SUPERSYMMETRY AND DUALITY 381
exp [iQp Jrworld - volume A p+1] = exp Q [i p ! AJlo ... JlpdxJlO 1\ ... 1\ dx JlP ] ,
(14.5)
382 ELIAS KIRITSIS
<I> 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).
SUPERSYMMETRY AND DUALITY 383
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
I
X min = (xi x~0)
0 (15.3)
SUPERSYMMETRY AND DUALITY 385
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
(16.2)
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
SUPERSYMMETRY AND DUALITY 387
(16.4)
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
SUPERSYMMETRY AND DUALITY 389
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)
L L
00 00
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
(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
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.
SUPERSYMMETRY AND DUALITY 393
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.
SUPERSYMMETRY AND DUALITY 395
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.
© 2000 Kluwer Academic Publishers.
398 MIKHAIL SHAPOSHNIKOV
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
400 MIKHAIL SHAPOSHNIKOV
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.
402 MIKHAIL SHAPOSHNIKOV
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
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
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)
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
0.5 V( lj»
0.25
-0.25
-0.5
-0.75
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
406 MIKHAIL SHAPOSHNIKOV
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
408 MIKHAIL SHAPOSHNIKOV
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
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
412 MIKHAIL SHAPOSHNIKOV
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
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
414 MIKHAIL SHAPOSHNIKOV
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
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
J-J Aubert et al. (eds.). Particle Physics: Ideas and Recent Developments. 417-510.
© 2000 Kluwer Academic Publishers.
418 PIERRE SALATI
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
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.
ds = dr (3)
~
(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)
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
(10)
Starting from the Robertson-Walker metric, readily derive the affine con-
nections (4), (5), (6) and (7).
COSMOLOGY AND DARK MATTER 421
z
N
M
r e
0Y--___------,+-_y
s
The flat case corresponds to k = O.
:0
I
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)
(16)
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)
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)
(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.
(22)
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)
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 '
COSMOLOGY AND DARK MATTER 425
2 87r G 0 k 87r G 0
Ho = --P --Pc (26)
3 ao
2 3
(28)
(29)
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 .
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
.......................................
Positive Pressure
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
Negative Pressure
v(.x) 1
x
Deceleration Acceleration
Acce .ration
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
COSMOLOGY AND DARK MATTER 429
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.
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
l _ { oWe } oh
- 47r a5 rr 1
Oto (1 + z) ,
(35)
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) ,
to = t to
0 dt -
-
H- I
0
t+oo
0 (1 + z)
dz
JF(z)
(39)
(40)
Problem nO 8 - Level [1] : Show that the right-hand side term of Eq. (40)
may be expressed as
h
rl k = 0 (fiat)
sinh- I (Fkrl) k < 0 (hyperbolic)
COSMOLOGY AND DARK MATTER 431
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)
d = ~(l+z) ~Iklr
L Ho JlTIKT V11\;1 I , (44)
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.
(46)
432 PIERRE SALATI
(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.~ __~ ~ _
.~
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] .
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)
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
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
COSMOLOGY AND DARK MATTER 435
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
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
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 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)
(58)
(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
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
t C
aatf IK = 1
x2 ox
a {x 4 ( ox
af + f + f 2) } , (66)
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)
y = J {Te - T,} dt .
Te tc
(69)
t e, -_ -3 {me
--
c} -
x;l- , (70)
4 O"T P,
mkT}3/2
n = 98 { 211" 1i 2 exp {(fJ. - mc2 ) / k T} (71)
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)
(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
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)
COSMOLOGY AND DARK MATTER 443
oT = _ ~ 1 <I> (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) <I>
(86)
T - (3a + 3) e'
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
II
~ 20
(87)
(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)
2
T = 30 1"(3) mBc (90)
7r4 .., 'T/ k '
-8p + \7.
-
(pv) o, (91 )
at
446 PIERRE SALATI
Problem nO 19 - Level [2] : Show that the density fluctuations follow the
linearized equations
COSMOLOGY AND DARK MATTER 447
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 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
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)
(108)
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
COSMOLOGY AND DARK MATTER 449
(111)
(112)
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)
(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
0.6
0 .6
0 .4
0.2
BOOMERANG/NA
+COBE
0.0 --L-..-.
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.
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] .
~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.
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)
~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.
'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
(J ~ 220 kms-
L
1 { Lo '
}1/4 (125)
(126)
(127)
(128)
o. (130)
We will assume that the halo is a sphere. Integrating out the velocity if
leads to the density profile
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<I>}
~ <I> = r2 dr r dr = 4 1f G p( r ) . (132)
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
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.
This is actually the method which Zwicky followed when he estimated the
dynamical mass contained in the Coma cluster.
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
(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)
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.
(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.
COSMOLOGY AND DARK MATTER 459
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
(149)
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
That element is a sensitive baryometer and allows to probe for OB. The
previous measurement translates into 7]10 ~ 5.5 and
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
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)
hoo = 2 <P
462 PIERRE SALATI
(161)
where the scalar potential cI> and the vector potential V which the system
generates are defined as
V(M, t)
= _G r pv(P, t') dTp (162)
c3 } distribution r
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
(166)
COSMOLOGY AND DARK MATTER 463
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
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.
n=1-2<1>+4V·e. (170)
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.
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)
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'
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) .
The vector (is horizontal and points from P towards the axis Oz
(177)
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)
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.
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
(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)
(188)
(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)
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.
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
_ 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)
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.
(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
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.
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 '
• 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
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)
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
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..
(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
= 47rGp (210)
COSMOLOGY AND DARK MATTER 475
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.
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)
(218)
COSMOLOGY AND DARK MATTER 477
(219)
T
T
= ~ = {8 G
a 3
1f
geff
(T) 1f2
15
T4} 1/2 .
(220)
(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.
< <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)
COSMOLOGY AND DARK MATTER 479
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.
Pv
°= 3({3)
2 7[2
T03
v
M '" 110k V
- e cm
-3 {~}
1 ke V ' (233)
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
COSMOLOGY AND DARK MATTER 481
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.
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.
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 ,
\
\
,
\
\
,,
\
\
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
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
COSMOLOGY AND DARK MATTER 483
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)
(241)
(242)
- - 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
df A
--2
_
-
{g5
--3 geff
-1/2
< aan V > M Mp dx (246)
fA 87f
(247)
- 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 >
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]
(249)
(250)
COSMOLOGY AND DARK MATTER 487
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 .
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 >
(253)
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
(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)
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)
(260)
that are related to the comoving wave vector k through the spectrum [35]
(261)
P(k) oc kn . (262)
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
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.
(270)
COSMOLOGY AND DARK MATTER 495
where mlOO = mx/lOO GeV and derive the corresponding spike radius
1/3 -1/3
a spike "'0.13pcA m lOO . (271)
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 '
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.
COSMOLOGY AND DARK MATTER 497
-
-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
(280)
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
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
COSMOLOGY AND DARK MATTER 499
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] .
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
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
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
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
COSMOLOGY AND DARK MATTER 503
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
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.
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.
COSMOLOGY AND DARK MATTER 505
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 <I>v varies therefore as
(286)
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
COSMOLOGY AND DARK MATTER 507
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.
<Pv = -BvNa
-2- = (-3
10 v s -1 cm -2) Am -1
lOO . (289)
47ff0
(290)
-38 2 ( Ev ) (291)
av-tJl. = 10 cm 1 GeV .
508 PIERRE SALATI
(292)
(293)
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).
COSMOLOGY AND DARK MATTER 509
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
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