discrete symmetries and proton decay in the … a isto e necess´ ario estudar nova f´ ´ısica...
TRANSCRIPT
Discrete Symmetries and Proton Decay in the Adjoint SU(5)Model
João Duarte Cardoso Texugo de Sousa
Thesis to obtain the Master of Science Degree in
Engineering Physics
Supervisor(s): Prof. Doutor David Emanuel da Costa
Examination Committee
Chairperson: Prof. Doutor Jorge Manuel Rodrigues Crispim RomãoSupervisor: Prof. Doutor David Emanuel da CostaMembers of the Committee: Prof. Doutor Filipe Rafael Joaquim
Prof. Doutor Palash Baran Pal
May 2015
Acknowledgments
First of all, I would like to thank my supervisor, professor David Emmanuel-Costa for his assistance and
support throughout this work. He was always available to help with the problems I encountered and
made frequent suggestions on how I could improve this thesis. I would also like to thank my friends who
were kind enough to provide me tips in programming-related issues and to debate certain aspects of
what I was doing. Finally, I also have to thank my family for giving me the means to complete this course
and always being there for me when I needed.
My experience in Instituto Superior Tecnico was a pleasant one, as I feel I was given all the necessary
means to learn about Physics. The hard work that was required helped me grow as a person and left
me prepared to deal with my future challenges in the academic world or other environment.
v
Resumo
O Modelo Padrao foi um marco importante na historia da Fısica, tendo obtido varios importantes suces-
sos, mas ja ha muito se sabe que esta teoria nao descreve correctamente toda a fenomenologia das
interaccoes entre partıculas elementares. Devido a isto e necessario estudar nova Fısica para alem
deste modelo.
Uma via interessante que pode ser seguida e a das Teorias de Grande Unificacao, nas quais um
dos principais problemas do Modelo Padrao, a ausencia de quantizacao das cargas electricas, e natu-
ralmente resolvido. Apesar disto e de outras vantagens que estas novas teorias oferecem, elas tambem
introduzem novos problemas, sendo o mais notavel dos quais, possivelmente, o decaimento do protao.
Nunca se observou um evento desta natureza, o que esta de acordo com a previsao do Modelo Padrao,
no qual este processo e proıbido devido a uma simetria B − L exacta. Por outro lado, nas Teorias de
Grande Unificacao surgem novos bosoes que podem mediar este decaimento, pelo que a viabilidade
destas depende de se encontrar uma forma de suprimir este processo.
Neste trabalho vai-se estudar a possibilidade de alcancar este objectivo, no contexto do modelo
SU(5) Adjunto, utilizando simetrias discretas. Apos esta analise vai-se tambem verificar se o modelo
resultante e compatıvel com os dados experimentais referentes as interaccoes electrofracas no limite
de baixas energias.
Palavras-chave: Teorias de Grande Unificacao, Decaimento do Protao, SU(5) Adjunto,
Simetrias Discretas
vii
Abstract
The Standard Model was an important milestone in the History of Physics, with several important suc-
cesses, but it has been known for long that it does not correctly describe all aspects related to the
phenomenology of interactions between elementary particles. As a consequence, it is necessary to
study new Physics beyond this model.
An interesting approach that may be taken is that of Grand Unified Theories, in which one of the
Standard Model’s main problems, absence of electric charge quantization, is naturally solved. In spite
of this and the other advantages these new theories offer, they also introduce new problems, the most
important of which being, arguably, proton decay. No such event has ever been observed and this fact is
in agreement with the Standard Model’s predictions, as the process is forbidden due to an exact B − L
symmetry. On the other hand, Grand Unified Theories imply new bosons that can mediate these decays
and, because of this, their viability depends on weather or not one can find ways of suppressing those
processes.
In this work we will investigate the possibility of achieving this, in the context of the Adjoint SU(5)
model, using discrete symmetries. After this analysis we will also check weather the resulting model is
consistent with experimental data related to electroweak interactions in the low-energy limit.
Keywords: Grand Unification Theories, Proton Decay, Adjoint SU(5), Discrete Symmetries
ix
Contents
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Resumo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
1 Introduction 1
2 The Standard Model of Particle Physics 5
2.1 Gauge group and Higgs mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Fermion masses and Mixings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Gauge Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3 SU(5) based Grand Unified Theories 23
3.1 SM shortcomings and justification for GUTs . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Minimal SU(5) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 SU(5) Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.1 Non-Renormalizable Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3.2 Renormalizable Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3.3 Seesaw Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4 Discrete Symmetries and Proton Decay 41
4.1 Adjoint-SU(5) Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.2 Proton Decay in the Adjoint SU(5) model . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.3 Conditions for Proton Decay Suppression and Discrete Symmetries . . . . . . . . . . . . 46
5 Adjoint SU(5) with Discrete Symmetry Results 51
5.1 Criteria for a Realistic Adjoint SU(5)×ZN theory . . . . . . . . . . . . . . . . . . . . . . . . 51
5.1.1 Low-Energy Neutrino Phenomenology . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.1.2 Discrete Gauge Symmetry Anomalies . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.1.3 Unification Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2 Study of two particular solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
xi
5.2.1 Z=8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.2.2 Z=7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6 Conclusions 69
Bibliography 75
A Renormalization Group Equations and complementary data 77
B Renormalization Group Equations and complementary data 79
B.1 Generalized Gell-Mann Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
B.2 Matter and Higgs Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
B.3 Scalar Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
C Discrete Symmetries that suppress Proton Decay and resulting Mass Matrices 85
C.1 Results for ZN × ZM symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
C.2 Results for continuous symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
xii
List of Tables
2.1 SM fermions and their GSM quantum numbers. . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1 New minimal SU(5) fields that contribute to the B-test and their influence. . . . . . . . . . 32
3.2 Contributions to the B-test from 45H fields. . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1 Sufficient conditions for proton decay suppression in the T1 mediated case. . . . . . . . . 47
4.2 Sufficient conditions for proton decay suppression in the T2 mediated case. . . . . . . . . 48
4.3 Sufficient conditions for proton decay suppression in the ∆−1/3 mediated case. . . . . . . 49
4.4 Sufficient conditions for proton decay suppression in the ∆2/3 mediated case. . . . . . . . 49
4.5 The first column indicates the number of Mu zeros and the first row indicates the number
of Md zeros. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.6 The first column indicates the number of Mu zeros and the first row indicates the number
of Md zeros. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.1 Relevant contributions to the B-test in the Adjoint SU(5)× ZN model. . . . . . . . . . . . 56
xiii
List of Figures
2.1 Scalar potential with µ2 > 0 (left) and µ2 < 0 (right) . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Lowest order contributions to Tµνλ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Lowest order contributions to Tµν . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4 Triangle diagrams with vertices vector-vector-axial for non-Abelian gauges. . . . . . . . . 19
3.1 SM gauge couplings running. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2 Diagram associated with th d=5 Weinberg operator. . . . . . . . . . . . . . . . . . . . . . 36
3.3 Feynman diagram representing the exchange of heavy particles that generates type I
seesaw. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.4 Feynman diagram representing the exchange of heavy particles that generates type II
seesaw. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.5 Feynman diagram representing the exchange of heavy particles that generates type III
seesaw. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.1 Diagrams associated with T1 mediated proton decay. Some diagrams represent several
processes as Q and L can correspond to different particles. . . . . . . . . . . . . . . . . . 47
4.2 Diagrams associated with T2 mediated proton decay. . . . . . . . . . . . . . . . . . . . . . 48
4.3 Diagrams associated with ∆−1/3 mediated proton decay. . . . . . . . . . . . . . . . . . . . 48
4.4 Diagrams associated with ∆2/3 mediated proton decay. . . . . . . . . . . . . . . . . . . . 49
5.1 Correlation plots involving VCKM moduli, specifically, Vub vs. Vcb. The red dots represent
this work’s results, the blue ones represent all the possibilities. . . . . . . . . . . . . . . . 61
5.2 Correlation plots involving VCKM moduli, specifically, Vus vs. Vcb. The red dots represent
this work’s results, the blue ones represent all the possibilities. . . . . . . . . . . . . . . . 61
5.3 Correlation plots involving VCKM moduli, specifically, Vus vs. Vub. The red dots represent
this work’s results, the blue ones represent all the possibilities. . . . . . . . . . . . . . . . 62
5.4 Correlation plots involving VCKM angles, specifically, γ vs. sin 2β . The red dots represent
this work’s results, the blue ones represent all the possibilities. . . . . . . . . . . . . . . . 62
5.5 Correlation plots involving VCKM angles, specifically, J vs. sin 2β. The red dots represent
this work’s results, the blue ones represent all the possibilities. . . . . . . . . . . . . . . . 63
xv
List of Acronyms
CKM Cabibbo-Kobayashi-Maskawa
GUT Grand Unified Theory
PMNS Pontecorvo-Maki-Nakagawa-Sakata
QFT Quantum Field Theory
RGE Renormalization Group Equations
SM Standard Model
SSB Spontaneous Symmetry Breaking
vev Vacuum Expectation Value
xvii
Chapter 1
Introduction
In the beginning of the 20th century there were two paradigm shifts in Physics, Quantum Mechanics and
Relativity. Amidst the many new discoveries made at this time, it was found that electromagnetic interac-
tions are very accurately described by a relativistic Quantum Field Theory (QFT), specifically, Quantum
Electrodynamics [1]. This theory’s Lagrangian is invariant under local transformations belonging to an
Abelian group, so it is also a gauge theory. In 1954, Chen Ning Yang and Robert Mills [2] generalized
this to non-Abelian groups while trying to explain strong interactions and, for this reason, gauge theo-
ries seeking to describe elementary particle interactions using any compact, semi-simple Lie group are
known as Yang-Mills theories.
These theories were, however, abandoned at an initial stage, because the gauge symmetry would
not allow fermion and gauge boson masses, which was contrary to experimental data. In 1964, Pe-
ter Higgs [3], Robert Brout and Francois Englert [4] and Tom Kibble, Gerald Guralnik and C. R. Ha-
gen [5] proposed, independently, a mechanism involving spontaneous symmetry breaking in the context
of gauge theories that would lead to mass terms for certain particle fields. The result became known as
the Higgs mechanism and it requires a scalar particle that acquires a vacuum expectation value. Using
this procedure, Weinberg and Salam published papers, in 1967 [6] and 1968 [7], respectively, in which
they unified electromagnetic and weak interactions in a Yang-Mills theory, first suggested by Sheldon
Glashow in 1961 [8], where the differences between these interactions were attributed to a spontaneous
breakdown of gauge symmetry. Since not much was known about the renormalizability of this theory it
did not attract much attention until some important results were later proved.
These results involve the quantization and renormalization of gauge theories. Regarding the first
aspect, significant progress was achieved by scientists like Feynman, DeWit, Mandelstam, Fadeev and
Popov. As for the second aspect, detailed studies of the simplest field theory with spontaneous symme-
try breaking revealed that this phenomenon does not affect the divergences of the theory and in 1971 G.
’t Hooft [9] demonstrated that Yang-Mills theories with spontaneous symmetry breaking are renormal-
izable. All these advances enabled physicists to obtain a renormalizable Yang-Mills theory describing
strong, weak and electromagnetic interactions in 1973-74 after the strong interactions were added to
Weinberg and Salam’s electroweak model.
1
The resulting theory is known as the Standard Model of Particle Physics (SM) and it had some re-
markable successes, including the prediction of gauge bosons, that were first detected in 1979 [10]
(namely the gluons, mediators of strong interactions). This added legitimacy to the SM and gauge the-
ories in general, as did the later (1983) discovery of W and Z bosons [11–13]. When we also take
into account the remarkable quality in describing elementary interactions (gravity is not being consid-
ered)and the recent detection of Higgs bosons [14, 15]), it is clear that the SM is one of the greatest
achievements in Physics. Despite this, we know that the SM is neither a complete nor definitive theory
since it possesses certain insufficiencies such as no quantization of gravitational interactions, absence
of neutrino masses, no viable candidate for dark matter, no quantization of the electric charge and hi-
erarchy problems. Solving these issues requires Physics beyond the SM and extensive work has been
done in this context. One class of theories attempting to provide a better explanation of elementary
interactions are the Grand Unified Theories, that have electric charge quantization as the main motiva-
tion, as aspects like the cancellation of proton and electron electric charges are very important for the
existence of the macroscopic world as we know it. This is achieved by having a gauge group larger than
the SM one that is also simple or a direct product of identical simple groups while embedding the SM
group and it undergoes spontaneous symmetry breaking in such a way that this last group is returned
in the low-energy limit. A relevant consequence is that the SM’s three gauge couplings (one for every
type of interaction) are replaced by a single one that is observed above a certain energy scale and is
split into the three SM couplings at low-energy scales after spontaneous symmetry breaking (hence the
name of this kind of theories).
The simplest GUTs are based on the SU(5) group and the simplest of these models is called minimal
SU(5). It was introduced in 1974 by Howard Georgi and Sheldon Glashow [16], having the merit of
providing electric charge quantization from theoretical aspects, but many other problems remain while
new ones, like proton decay and wrong mass predictions, arise. This may not seem very promising,
however, the runing of SM gauge couplings, which almost unify at a certain scale, the aforementioned
accomplishments and the fact that most of these issues can be solved without radically changing the
theory make SU(5) GUTs a well-motivated framework and a good starting point for building a flavour
symmetry. Many models may be obtained by altering the minimal scenario and the background for this
thesis is one of these, the Adjoint-SU(5) [17]. Its advantages and disadvantages will be debated later
but the main focus is on proton decay. In fact, this problem affects GUTs in general as they imply the
existence of new fields that can mediate these decays but no such event has been detected thus far
and, as a consequence, the new fields in question have their masses strongly constrained. Therefore,
it is of interest to study possibilities of eliminating this issue. One interesting way of doing so is by
means of adding discrete symmetries to the gauge group and the main purpose of this thesis is to study
the symmetries that may successfully be used and weather the models obtained are consistent with
experimental data.
This work starts with a brief review of the Standard Model of Particle Physics, especially of its elec-
troweak sector as it contains the phenomenology that is more relevant for the rest of the thesis. Gauge
anomalies are also mentioned as they spoil a theory’s renormalizability and must, therefore, be absent
2
in any realistic model. Chapter 3 is dedicated to Grand Unified Theories and their general features are
introduced in a discussion of the minimal SU(5) model. This discussion reveals that the minimal setup
has a considerable number of problems so several pertinent extensions are presented to address them.
Of particular importance are the see-saw mechanisms that allow neutrinos to acquire mass. In Chapter
4 the Adjoint model is introduced and we study the proton decay processes that may take place in this
context. A discrete symmetry is introduced with the purpose of loosening constraints related to this
issue while reducing the arbitrariness in the Yukawa sector. In Chapter 5 we attempt to build realistic
theories without tree-level scalar-mediated proton decays by choosing particular discrete symmetries
and associated charges. Finally, this work’s conclusions are summarized in Chapter 6.
3
Chapter 2
The Standard Model of Particle
Physics
The SM is, without a doubt, an important milestone in particle physics. In spite of its shortcomings, it
gives a very good description of low-energy interactions, which is why many models of greater complexity
return it (with a few possible changes) as an effective theory in this limit. Besides, some of its main
”ingredients”, such as a gauge group and the Higgs mechanism are present in a wide variety of more
sophisticated gauge theories, so it is very important to understand them. In this Chapter we go over
some of the SM’s main features, with emphasis on the electroweak sector. Afterwards, we make a
simplified analysis of gauge anomalies in order find out which conditions must be verified for them to be
absent and prove that the SM is anomaly-free.
2.1 Gauge group and Higgs mechanism
The Standard Model is a relativistic Quantum Field Theory. As happens with this kind of theories, be
they classical or quantum in nature, the main quantity is the Lagrangian L, from which we can obtain the
field’s equations of motion by applying the principle of stationary action:
δS = δ
∫d4xL = 0⇔
∫d4x
[∂L∂φi
δφi +∂L
∂(∂µφi)δ(∂µφi)
]=
∫d4x
[∂L∂φi− ∂µ
∂L∂(∂µφi)
]δφi
⇒ ∂L∂φi− ∂µ
∂L∂(∂µφi)
= 0.
(2.1)
These are known as the Euler-Lagrange equations.
More specifically, the SM is a gauge theory, which means that its Lagrangian is invariant for a certain
group of local transformations. That group is
GSM = SU(3)C × SU(2)L × U(1)Y (2.2)
where SU(3)C is associated with strong interactions, SU(2)L is associated with weak interactions and
5
U(1)Y is the hypercharge group. Local transformations have an explicit dependence on the space-
time coordinates so terms with gradients are not invariant. For a non-Abelian group (like groups from
the SU(n) family) this problem can be solved by adding vector bosons to the theory and replacing the
gradients with covariant derivatives using the following prescription (minimal coupling) [18]∂µ → Dµ =
∂µ − igL.Aµ(x). In the previous equation, g is the coupling constant, Li are matrix representations of
the group generators and Aiµ(x) are the vector boson fields. This is easily generalized to Abelian cases
by replacing the Li with a constant. It is important to note that one vector boson is added for every
generator of the gauge group.
From this procedure of minimal coupling and (2.2) we can see that, in the SM, covariant derivatives
may be written as
Dµ = ∂µ − igS8∑a=1
Gaµλa
2− ig
3∑a=1
W aµ
σa
2− igY
Y
2Bµ, (2.3)
where the second term on the right corresponds to the SU(3)C group, whose generators are represented
by the λa Gell-Mann matrices (shown in Appendix B), the third term corresponds to the SU(2)L group,
whose generators are represented by the σa Pauli matrices and the last term corresponds to U(1)Y ,
represented by the constant Y. There are 12 vector bosons, 8 Gaµ, 3 W aµ and Bµ.
The interactions between gauge bosons and other fields are given by the kinetic terms of the other
fields, with partial derivatives replaced by covariant ones. It is still necessary to add the kinetic terms
for the gauge bosons (terms that only include these fields or their derivatives). In the Abelian case the
field-strength tensor is defined as
Fµν = ∂µAν − ∂νAµ (2.4)
and the gauge-kinetic term to be added to the Lagrangian is
− 1
4FµνFµν . (2.5)
In the non-Abelian case the field-strength tensor is
F aµν = ∂µAaν − ∂nuAaµ − gfabcAbµAnuc, (2.6)
where fabc are the structure constants of the group (for a definition of structure constants and a review
of Lie Algebras see [19]). The gauge-kinetic term has the same form as (2.5). Considering this, the
terms that must be added to the SM Lagrangian are
Lgauge−kinetic = −1
4BµνBµν −
1
4W aµνW a
µν −1
4GaµνGaµν . (2.7)
At this point we note that mass terms for the gauge bosons of the form
m2AµAµ (2.8)
would explicitly break the gauge symmetry. It has been experimentally observed that some of these
6
bosons have mass, namely, the ones associated with weak interactions, and it is known that the cor-
responding isospin symmetry is broken. We could consider explicitly breaking the symmetry by adding
these mass terms, but the gauge boson masses would be arbitrary parameters of the model. Instead,
we look for another way of breaking the symmetry that offers more predictivity, namely, Spontaneous
Symmetry Breaking. Furthermore, if we choose the path of explicit symmetry breaking, certain Feyn-
man diagrams have ”worse” divergences than in the case with Spontaneous Symmetry Breaking and, in
that context, unitarity is lost, while renormalizability itself would, in general, also be lost [20].
How can a symmetry be broken in a non-explicit way and how can the bosons related to the broken
symmetry acquire mass in that scenario? The answer is provided by the so-called Higgs mechanism [3–
5]. A scalar field φ is postulated to exist. This field is an SU(2)L doublet that can be written as
φ =
φ+
φ0
. (2.9)
The most general renormalizable potential for φ is
V (φ†φ) = µ2(φ†φ) + λ(φ†φ)2, (2.10)
where µ and λ are constants. If λ < 0 the field oscillations are unbounded so λ is taken to be positive.
On the other hand, µ2 can be positive or negative. When it is considered that µ2 > 0, the minimum of
the potential is 0 with 〈φ〉0 = 0. When µ2 is considered to be negative the potential changes and its
minimum is no longer 0, as shown in Figure 2.1. The minimum occurs for 〈φ〉0 = −µ2
λ = v2 now, where
v is a real and positive constant. From this point on only the case with µ2 < 0 is considered for reasons
that will become clear soon.
Figure 2.1: Scalar potential with µ2 > 0 (left) and µ2 < 0 (right)
Taking advantage of the freedom to perform SU(2)L rotations the vacuum state of φ can be parametrized
as
〈φ〉0 =1√2
0
v
. (2.11)
Contrarily to the case where 〈φ〉0=0, this vacuum does not preserve the SU(2)L×U(1)Y gauge symmetry
since
eiαk〈φ〉0 ≈ (1 + iαk)〈φ〉0 6= 〈φ〉0, (2.12)
7
where the k indexes refer to SU(2)L × U(1)Y generators. This means that, in general, electroweak
gauge group transformations don’t leave the vacuum invariant and, consequently, there is a symmetry
breaking. This phenomenon is usually referred to as Spontaneous Symmetry Breaking (SSB) because
the Lagrangian itself has a certain gauge symmetry but the physical system as a whole does not. There
is still an U(1) subgroup of the electroweak gauge group under which the theory is invariant. Defining
Q =σ3
2+ Y (2.13)
and assigning φ an hypercharge of 12 the result of acting with Q upon the vacuum of the scalar field is
Q〈φ〉0 =
1 0
0 0
0
v
=
0
0
. (2.14)
This Q leaves the vacuum invariant and it is the generator of the residual U(1) gauge group.
From the previous discussion one can see that 3 of the 4 generators of the electroweak gauge group
are broken so, according to Goldstone’s Theorem [21], there should be 3 massless Goldstone bosons
in the theory. At this point, two problems arise when comparing the theory described so far with experi-
mental results: no Goldstone bosons have been observed and three gauge bosons are massive. These
issues are related and have a common explanation, the so-called Higgs mechanism. The fact that the
Lagrangian has gauge invariance allows one to make certain local transformations without changing the
physics of the model. It is possible, using one of those transformations, to make the Goldstone bosons
”disappear”. These bosons are ”gauged away” and the gauge in which this happens is called unitary
gauge. Being more rigorous, the Goldstone bosons do not completely vanish, they are incorporated as
degrees of freedom of gauge boson fields that, as a consequence, become massive [18].
It has already been stated that mass terms for the gauge bosons can’t be included in the gauge-
kinetic part of the Lagrangian. Instead, they arise from the kinetic term of the Higgs field. This field, in
the unitary gauge, can be parametrized as
φ =
0
(v +H)/√
2
, (2.15)
where H is a real scalar field. Looking at the scalar potential in (2.10) we note, using (2.15), that there
are small oscillations around the vev which correspond to a boson with a mass of
m2H = −2µ2 = 2v2λ. (2.16)
With this parametrization the kinetic term of φ is
(Dµφ)†(Dµφ) =1
2∂µH∂
µH + g2 (v +H)2
4
1√2
(A1µ + iA2
µ)1√2
(A1µ − iA2µ)
+(v +H)2
8(g2A3
µA3µ − ggYA3
µBµ − ggYA3
µBµ − g2
YBµBµ).
(2.17)
8
The A1µ, A2
µ, A3µ and Bµ fields are not mass eigenstates so they are not physical fields. It is, then, useful
to introduce new fields that are mass eigenstates. Regarding the second term on the right-hand side of
(2.17), the new fields are defined as
W±µ =A1µ ∓ iA2
µ√2
(2.18)
and this term becomes
g2 v2
4W−µ W
+µ, (2.19)
from which it can be seen that the new fields have a mass of
MW =
√1
4g2v2 =
1
2gv. (2.20)
Ignoring the H field (because at this point the focus is on the boson masses), the third term on the right
of (2.17) may be written as
v2
8(A3
µBµ)
g2 −ggY−ggY g2
Y
A3µ
Bµ
. (2.21)
Finding the mass eigenstates is, in this case, equivalent to diagonalizing the mass matrix in the previous
equation. This is done by defining the new fields as the ones that give the old fields when rotated by a
certain angle (Weinberg angle):A3µ
Bµ
=
cos θW sin θW
− sin θW cos θW
ZµAµ
, (2.22)
with
cos θW =g√
g2 + g2Y
, sin θW =gY√g2 + g2
Y
. (2.23)
Using (2.22) in (2.21) the mass term becomes
v2
8(ZµAµ)
g2 + g2Y 0
0 0
ZµAµ
, (2.24)
from which it can be seen that Zµ has a mass of MZ = v2
√g2 + g2
Y and Aµ has no mass. The boson
associated with electroweak interactions that remains massless after SSB is the photon so Aµ is iden-
tified as this field [22]. Breaking the gauge symmetry through SSB we have obtained all gauge boson
masses in terms of other parameters of the model.
It is convenient to rewrite the covariant derivative given by (2.3) in terms of the physical fields. Defin-
ing
T± =σ1/2± iσ2/2√
2(2.25)
one gets:
Dµ = ∂µ − ig√2
(W+µ T
+ +W−µ T−)− i 1√
g2g2Y
Zµ(g2σ3/2− g2Y Y )− i ggY√
g2 + g2Y
Aµ(σ3/2 + Y ). (2.26)
9
The last term involves the photon, therefore, it corresponds to the electromagnetic interaction. Knowing
this, one can conclude thatggY√g2 + g2
Y
= g sin θW = e (2.27)
and σ3/2 + Y gives the electric charge. The previously defined Q is, then, the electric charge operator
and the U(1) gauge group it generates corresponds to electromagnetism. This gauge symmetry is not
broken in the SM and the electric charge is conserved, in accordance with all experimental data collected
so far. Making appropriate substitutions, (2.26) can be written as
Dµ = ∂µ − ig√2
(W+µ T
+ +W−µ T−)− i g
cos θWZµ
(σ3
2− sin2 θWQ
)− ieAµQ. (2.28)
2.2 Fermion masses and Mixings
Other than the gauge bosons and Higgs fields, the SM also includes another type of particles, elemen-
tary fermions. These are Dirac particles with a spin of 1/2 and form most of the matter we see around
us (most of the bound states found in nature are constituted by them). Fermions can be classified as
quarks, which experience all four fundamental interactions in nature, or leptons, that don’t have strong
interactions. These particles, in the SM, appear as left-handed or right-handed fields where
ΨL =1− γ5
2Ψ, ΨR =
1 + γ5
2Ψ. (2.29)
The left-handed components are SU(2)L doublets while the right-handed ones are singlets under this
group. Regarding the SU(3)C group, quarks form triplets while the leptons are singlets, as one would
expect given the fact that they don’t experience strong interactions. Finally, the hypercharge quantum
numbers are assigned in order to get the correct (experimentally observed) electric charge for the fields
from (2.14). The quantum number of the fermion fields in relation to GSM are displayed in Table 1.
Quark Fields Quantum Numbers Lepton Fields Quantum Numbers
qL =
(uLdL
)(3,2,1/6) LL =
(νe−
)(1,2,−1/2)
uR (3,1,2/3) eR (1,1,−1)
dR (3,1,−1/3)
Table 2.1: SM fermions and their GSM quantum numbers.
10
It is known that fermions are massive but mass terms like
LFermionMass = −m(ψRψL + ψLψR) (2.30)
cannot be added to the Lagrangian because they would explicitly break the electroweak gauge symme-
try. This is due to the fact that right-handed and left-handed fields belong to different representations of
the electroweak gauge group. There must be some other way for the fermions to acquire mass in the
SM.
Using the scalar field φ it is possible to obtain SU(2)L ×U(1)Y -invariant terms that mix right-handed
and left-handed fermion fields:
LYukawa = −QiLY uij φujR − Q
iLY
dijφd
jR − L
iLY
eijφe
jR +H.c., (2.31)
where φ = iσ2φ∗ and the Y u,d,e are arbitrary complex 3 × 3 Yukawa matrices. When φ gets a vev of v
there is SSB and the following mass terms arise from the previous equation:
LMass = −uiLM iju u
jR − d
iLM
ijd d
jR − e
iLM
ije e
jR +H.c.. (2.32)
The fermion mass matrices, Mu, Md and Me are given by
Mu =1√2vY u, Md =
1√2vY d, Me =
1√2vY e. (2.33)
It should be noted that, since there are no right-handed neutrino fields in the SM, these particles do not
acquire mass in this way.
Since the Yukawa matrices are not, in general, diagonal, the fermion eigenstates of the electroweak
interactions are not mass eigenstates. These last eigenstates are obtained by diagonalizing the mass
matrices using the following bi-unitary transformations:
uL = UuLu′L, uR = UuRu
′R, (2.34)
dL = UdLd′L, dR = UdRd
′R, (2.35)
eL = UeLe′L, eR = UeRe
′R, (2.36)
νL = UeLν′L, (2.37)
where the Us are unitary matrices and the primed fields are the physical fields (mass eigenstates). After
these transformations are performed the masses become
Uu†L MuUuR = diag(mu,mc,mt) = Du, (2.38)
Ud†L MdUdR = diag(md,ms,mb) = Dd, (2.39)
11
Ue†L MeUeR = diag(me,mµ,mτ ) = De. (2.40)
It should be stated that the m values represent actual masses so they are real and positive.
Due to the arbitrariness of the Yukawa matrices, UuL is different from UdL and this leads to mixings
between quarks. To see how this happens it is convenient to analyse the interactions of fermions with
the gauge bosons. These interactions are given by the Dirac Lagrangian with the partial derivatives
replaced by covariant ones:
LInteractions = g(W+µ J
µ+W +W−µ J
µ−W + ZµJ
µZ) + eAµJ
µEM , (2.41)
where the J’s are currents that may be written as
Jµ+W =
1√2
(uLγµdL + νLγ
µeL), Jµ−W =1√2
(dLγµuL + eLγ
µνL), (2.42)
JµZ =1
cos θW
[uLγ
µ(1
2− 2
3sin2 θW )uL + uRγ
µ(−2
3sin2 θW )uR + dLγ
µ(−1
2+
1
3sin2 θW )dL+
+ dRγµ(
1
3sin2 θW )dR + νLγ
µ 1
2νL + eLγ
µ(−1
2+ sin2 θW )eL + eRγ
µ sin2 θW eR
],
(2.43)
JµEM =2
3uγµu− 1
3dγµd− eγµe. (2.44)
When fermions fields are rotated to the mass eigenstate basis, the left-handed components of the quarks
get mixed, as one can see in the positive charged weak current:
Jµ+W =
1√2
(u′LγµUu†L UdLd
′L + ν′Lγ
µUe†L UeLe′L). (2.45)
While Ue†L UeL is the identity and no mixing occurs on the lepton sector, Uu†L UdL = V is different form the
identity, it is a complex 3× 3 unitary matrix known as the CKM matrix [23,24], that may be written as
V =
Vud Vus Vub
Vcd Vcs Vcb
Vtd Vts Vtb
. (2.46)
Note that, since neutrinos are massless in the SM, we can change their basis using the same unitary
matrix as we do for left-handed charged leptons and, consequently, no lepton mixing occurs.
Some parameters of the VCKM are devoid of physical meaning, as we are free to rephase quark
fields through
uα = eiΨαu′α
dk = eiΨkd′k,(2.47)
where the Ψ are arbitrary phases. These transformations lead to
V ′αk = ei(Ψk−Ψα)Vαk (2.48)
12
and, with ng fermion generations, we can use this to eliminate 2ng−1 VCKM phases [25]. Usually, being
unitary, this matrix would contain n2g parameters but, considering the previous discussion, this number
becomes
Nparam = n2g − (2ng − 1) = (ng − 1)2. (2.49)
Out of these,
Nangle =1
2ng(ng − 1) (2.50)
may be identified as rotation (Euler) angles. The remaining
Nphase = Nparam −Nangle =1
2(ng − 1)(ng − 2) (2.51)
parameters correspond to physical phases. For ng = 3 we see that there is only one such phase
and it can be shown that this phase leads to CP violation [25]. The phenomenon of CP violation is
very interesting as it can provide (along with the assumption of departure from thermal equilibrium
and non-conservation of baryon number) an explanation for the baryon asymmetry in the observable
universe [26], but the CP violation predicted in the SM is not sufficient [27]. We could consider this a
problem, however, there are other possible sources of baryon asymmetry so we dismiss this issue and
CP violation will not be studied in the remaining chapters.
Later in this work we will want to compare the predicted VCKM with experimental results related to
quark mixings, so it is of interest to define rephasing-invariant quantities that may be measured through
appropriate experiments. The simplest possibility is given by
Uαi ≡ |Vαi|2, (2.52)
the moduli of the matrix elements. Another one is provided by invariant quartets, defined as
Qαiβj ≡ VαiVβjV ∗αjV ∗βi, (2.53)
where α 6= β and i 6= j. Since the VCKM matrix is unitary, its rows and columns must verify orthogonality
conditions. Considering, for example, the first and third columns we have
VudV∗ud + VcdV
∗cb + VtdV
∗tb = 0. (2.54)
We may interpret this equation as a triangle in the complex plane and, while rephasing of the quark fields
rotates the triangle, all internal angles and length of its sides remain invariant. Choosing a convention in
which VcdV ∗cb is real and negative, the inner angles of the triangle are defined by
α ≡ arg(−Qubtd),
β ≡ arg(−Qtbcd),
δ ≡ arg(−Qcbud).
(2.55)
13
The moduli in (2.52) and the α, β and δ are well measured physical quantities and we will use them
to test weather or not our predicted VCKM is consistent with experimental data. Note also that, by
definition, α+ β + γ = π mod 2π
2.3 Gauge Anomalies
The Lagrangian function plays an important role in the SM and QTF’s in general but, unlike what hap-
pens in classical theories, one must consider corrections to the interactions given by this function and
even possible non-perturbative effects beyond the Lagrangian (these latter are not relevant in the context
of this work). Radiative corrections are mentioned here because they can have important consequences
for the theory, such as violating symmetries of the underlying Lagrangian and jeopardizing its renormal-
izability.
In classical theories there is a correspondence between symmetries of the Lagrangian and con-
served charges given by Noether’s theorem. As for quantum field theories, the symmetry properties of
the Lagrangian lead to relations between Green’s functions, which are known as Ward Identities [28].
These identities have several applications but in this context the focus will be on their role in the renor-
malization programme of any theory with nontrivial symmetries. If the identities are violated by radiative
corrections then it is impossible to prove the renormalizability of the theory [9]. A particular case of
this will be analysed in this section, specifically the Adler-Bell-Jackiw anomaly [29, 30], named after the
researchers that discovered it in 1969-1970.
To begin with, currents may be classified according to their transformation properties as scalar (S),
vector (V), tensor (T), pseudoscalar (P) or axial vector (A):
S(x) = Ψ(x)Ψ(x), (2.56)
Vµ(x) = Ψ(x)γµΨ(x), (2.57)
Tµν(x) = Ψ(x)γµγνΨ(x), (2.58)
P (x) = Ψ(x)γ5Ψ(x), (2.59)
Aµ(x) = Ψ(x)γµγ5Ψ(x). (2.60)
Considering the following three-point functions of electrodynamics,
Tµνλ(k1, k2, q) = i
∫d4x1d
4x2 〈0 | T (Vµ(x1)Vν(x2)Aλ(0)) | 0〉 eik1·x1+ik2·x2 (2.61)
and
Tµν(k1, k2, q) = i
∫d4x1d
4x2 〈0 | T (Vµ(x1)Vν(x2)P (0)) | 0〉 eik1·x1+ik2·x2 , (2.62)
where q = k1 + k2, one may obtain the Ward identities relating Tµνλ and Tµν to check weather they are
14
verified. Using the divergences of Vµ and Aµ, which are calculated from the equation of motion
∂µVµ(x) = 0 (2.63)
∂µAµ(x) = 2imP (x), (2.64)
where m is the mass of Ψ, current-algebra techniques like
∂µx (T (Jµ(x)O(y))) = T (∂µJµ(x)O(y)) + [J0(x), O(y)] δ(x0 − y0) (2.65)
and with the knowledge that
[V0(x), A0(y)] δ(x0 − y0) = 0, (2.66)
we get the following vector and axial-vector Ward Identities [31]
kµaTµνλ = kνb Tµνλ = 0, (2.67)
qλTµνλ = 2mTµν . (2.68)
Figure 2.2: Lowest order contributions to Tµνλ.
The lowest order contributions to Tµνλ and Tµν come from the diagrams shown in Figures 2.2 and
2.3, respectively, and may be written as
Tµνλ = i
∫d4p
(2π)4(−1)
Tr
[i
�p−mγλγ5
i
(�p− �q)−mγν
i
(�p− �k1)−mγµ
]+
k1 ↔ k2
µ↔ ν
; (2.69)
Tµν = i
∫d4p
(2π)4(−1)
Tr
[i
�p−mγ5
i
(�p− �q)−mγν
i
(�p− �k1)−mγµ
]+
k1 ↔ k2
µ↔ ν
. (2.70)
Using
�qγ5 = γ5(�p− �q −m) + (�p−m)γ5 + 2mγ5 (2.71)
15
Figure 2.3: Lowest order contributions to Tµν .
we can obtain
qλTµνλ = 2mTµν + ∆(1)µν + ∆(2)
µν , (2.72)
with
∆(1)µν =
∫d4p
(2π)4Tr
{i
�p−mγ5γν
i
(�p− �k1)−mγµ −
i
(�p− �k2)−mγ5γν
i
(�p− �q)−mγµ
}(2.73)
and
∆(2)µν =
∫d4p
(2π)4Tr
{i
�p−mγ5γν
i
(�p− �k2)−mγµ −
i
(�p− �k1)−mγ5γν
i
(�p− �q)−mγµ
}. (2.74)
One can readily see that the Ward Identity in (2.68) is only respected if ∆(1)µν + ∆
(2)µν = 0. The two
integrals in (2.73) would cancel each other if one could make the shift p → p + k2 in the second term
inside the trace. Applying the same reasoning to (2.74) and performing the shift p→ p+k1 in the second
term inside the trace we see that this expression would also vanish. However, this procedure can only
be applied to convergent integrals, which constitutes a problem since these ∆s are linearly divergent.
Before studying the conditions in which (2.68) can be valid it is important to clarify this last point.
Starting with the one-dimension case it is easy to show that a shift of integration variable may be
impossible if the integral is divergent [32]. For an integration variable shift to be legitimate the quantity
given by
∆(a) =
∫ +∞
−∞dx [f(x+ a)− f(x)] (2.75)
must vanish. Expanding the integrand and taking into account the fact that, for linearly divergent inte-
grals, f(±∞) 6= 0, one gets
∆(a) = a [f(∞)− f(−∞)] , (2.76)
a surface term that is, in general, different from 0.
Generalizing (2.75) to an n−dimensional space, expanding the integrand, integrating over the surface
r = R →∞ and applying Gauss’s theorem, one can see that only the first term of the expansion, given
16
by
∆(a) = aτRτRf(R)Sn(R) (2.77)
remains, where RτR is the outward pointing unit normal field and Sn(R) is the surface area of the hyper-
sphere with radius R. In the four-dimensional Minkowski space, S3(R) = 2π2R3 so we get
∆(a) = aτ∫d4r∂τf(r) = 2iπ2aτ limR→∞R
2Rτf(r). (2.78)
The fact that Tµνλ is linearly divergent implies that this quantity is not uniquely defined. In (2.69)
the fermion line between the vector and axial-vector vertices carries momentum p but we may, instead,
assign it a momentum of p+ a where a is an arbitrary linear combination of k1 and k2:
a = αk1 + (α− β)k2. (2.79)
The ambiguity in the definition of Tµνλ may be measured through the difference between amplitudes in
∆µνλ = Tµνλ(a)− Tµνλ(0) =
= (−1)
∫d4p
(2π)4
{Tr
[1
(�p+ �a)−mγλγ5
1
(�p+ �a− �q)−mγν
1
(�p+ �a− �k1)−mγµ
]−
−Tr
[1
p−mγλγ5
1
(�p− �q)−mγν
1
(�p− �k1)−mγµ
]}+
k1 ↔ k2
µ↔ ν
≡≡ ∆
(1)µνλ + ∆
(2)µνλ,
(2.80)
where Tµνλ(a) is the shifted amplitude. Using (2.78) one gets
∆(1)µνλ = (−1)
∫d4p
(2π)4aτ
∂
∂pτtr
[1
�p−mγλγ5
1
(�p− �q)−mγν
1
(p− k1)−mγµ
]=
=−i2π2aτ
(2π)4limp→∞
p2pτ tr(γαγλγ5γβγνγδγµ)pαpβpδ/p6 =
=i2π2aσ(2π)4
limp→∞
pσpρ
p24iεµνλρ.
(2.81)
Considering the symmetric limit we replace pσpρ/p2 with gρσ/4 and we may write
∆(1)µνλ = ερµνλa
ρ/8π2. (2.82)
There is no need to compute ∆(2)µνλ explicitly as it only differs from ∆
(1)µνλ by the exchanges k1 ↔ k2 and
µ↔ ν. Bearing this in mind and combining (2.80), (2.82) and (2.79) we get
∆µνλ = ∆(1)µνλ + ∆
(2)µνλ =
β
8π2ερµνλ(k1 − k2)ρ, (2.83)
from which we conclude that the ambiguity in Tµνλ may be expressed in terms of the arbitrary parameter
17
β:
Tµνλ(a) = Tµνλ(0)− β
8π2εµνλρ(k1 − k2)ρ ≡ Tµνλ(β). (2.84)
A very important question arises at this point: is there a value of β for which the Ward Identities are
verified? Starting with the Axial Ward Identity (2.68), we can use (2.78) to evaluate the linearly divergent
terms in (2.72), which yields
∆(1)µν = − kτ2
(2π)4
∫d4p
∂
∂pτ(tr[(�p+m)γ5γν(�p−��k1 +m)γµ
](p2 −m2) [(p− k1)2 −m2]
) =
= − kτ2(2π)4
2iπ2 limp→∞
pτp2tr(γαγ5γνγβγµ)pαkβ1 =
=−1
8π2εµνσρk
σ1 k
ρ2
(2.85)
and
∆(2)µν = ∆(1)
µν . (2.86)
Therefore, from (2.84) and (2.72) we obtain
qλTλνλ(β) = 2mTµν(0)− 1− β4π2
εµνσρkσ1 k
ρ2 (2.87)
and it is clear that the Axial Ward Identity is verified only if β = 1.
Moving on to the Vector Ward Identity (2.67) we have
kµ1Tµνλ(0) =(−1)
∫d4p
(2π)4
{tr
[1
�p−mγλγ5
1
(�p− �q)−mγν
1
(�p−��k1)−mγµ
]+
+tr
[1
�p−mγλγ5
1
(�p− �q)−m��k1
1
(�p−��k2)−mγν
]}.
(2.88)
Using the relation
�k1 = (�p−m)−[(�p− �k1)−m
]=[(�p− �k2)−m
]− [(�p− �q)−m] (2.89)
we can write (2.88) as
kµ1Tµνλ(0) = (−1)
∫d4p
(2π)4tr
[γλγ5
1
(�p− �q)−mγν
1
(�p− �k1)−m− γλγ5
1
(�p− �k2)−mγν
1
�p−m
]. (2.90)
Like before, we can use (2.78) to evaluate the linearly divergent integrals in the previous expression,
which leads to
kµ1Tµνλ(0) =kτ1
(2π)4
∫d4p
∂
∂pτ(Tr[γλγ5(�p− �k2 +m)γν(�p+m)
][(p− k2)2 −m2] (p2 −m2)
) =
=kτ1
(2π)42iπ2 lim
p→∞
pτp2tr(γ5γλγαγνγβ)kα2 p
β =
=−1
8π2ελσνρk
ρ1kσ2 .
(2.91)
18
Finally, remembering (2.84) we get
kµ1Tµνλ(β) =(1 + β)
8π2ενλσρk
σ1 k
ρ2 . (2.92)
It is clear from this expression that the Vector Ward Identity is only verified when β = −1 and, more
importantly, the Ward Identities in (2.67) and (2.68) cannot be simultaneously respected for any value of
β. This means that at least one of the Ward Identities is violated. When a symmetry of the Lagrangian
is broken by a perturbative correction to the theory, as in the case being studied, we say we have an
anomaly.
Although it is already evident that we have an anomaly in this situation, the parameter β is still not
fixed, which constitutes a problem since it has physical consequences. Experimental results show that
the Vector Ward Identity is verified, therefore, β = −1 and the Axial Ward Identity is violated. The fact
that we were unable to obtain the value of β is related to the renormalization scheme used, which has no
physical meaning (it is just a mathematical device) and other renormalization procedures lead directly to
(2.67) being verified [33,34]. Considering, then, β = −1 the Axial Ward Identity becomes
qλTµνλ = 2mTµν −1
2π2εµνσρk
σ1 k
ρ2 (2.93)
and the axial-vector current divergence is modified to
∂λAλ(x) = 2imP (x) + (4π)−2εµνρσFµν(x)Fρσ(x). (2.94)
The anomaly studied until this point is a non-gauge Abelian chiral anomaly, so it does not threaten
renormalizability, instead it implies that some classically forbidden processes may occur.
Figure 2.4: Triangle diagrams with vertices vector-vector-axial for non-Abelian gauges.
As for gauge anomalies, they are the ones that jeopardize the renormalizability of a theory, so it is
important to study this phenomenon. We will take a look at non-Abelian gauge anomalies since we
work with non-Abelian gauge theories and generalization of the most important results to Abelian cases
is trivial. These anomalies are more complex than the ones treated thus far and a detailed analysis
19
is beyond the scope of this work. We can, however, perform certain computations to obtain a very
important result that remains valid in more accurate studies. Considering the diagrams in Fig. 2.4 with
non-Abelian vertices the amplitude in (2.69) is modified to
T abcµνλ = −i∫
d4p
(2π)4Tr
[i
�p−mγλγ5t
c i
�p− �q −mγνt
a i
�p− �k1 −mγµt
b
]+
k1 ↔ k2
µ↔ ν
a↔ b
. (2.95)
The gamma matrices commute with group generators so we can write
T abcµνλ = −i∫
d4p
(2π)4Tr
[i
�p−mγλγ5
i
�p− �q −mγν
i
�p− �k1 −mγµ
]Tr[tctatb
]+
k1 ↔ k2
µ↔ ν
a↔ b
. (2.96)
This modification in the amplitude leads to a change in the anomalous term given in (6.93):
Aabcµν =1
4π2εµναβk
α1 k
β2 Tr
[tctatb
]+
1
4π2ενµαβk
α2 k
β1 Tr
[tctbta
]=
=1
4π2εµναβk
α1 k
β2 Tr
[tctatb + tctbta
]=
=1
4π2εµναβk
α1 k
β2 Tr
[{ta, tb
}tc].
(2.97)
This was a grossly oversimplified calculation but it turns out that the anomalous terms are, in general,
proportional to Tr[{ta, tb
}tc]
[35], so, from now on, we may take the anomaly freedom condition to be
Tr[{ta, tb
}tc]
= 0. (2.98)
Using (2.98) and the facts that fermions contribute additively to anomalies, while left- and right-
handed fermions contribute with opposite signs, it is easy to check that the SM is anomaly-free. This
last aspect can be seen by noting that the charge conjugate of left-handed field is a right-handed field
and vice-versa and applying the charge conjugation operator to the generators in (2.98) we get the same
expression preceded by a minus sign. Recalling that, in the SM, the gauge group is a direct product of
three groups, the possible anomalies are associated with triangle diagrams that couple to these groups’s
gauge bosons. We can see, then, that 10 different anomalies may arise:[U(1)Y ]3, [SU(2)L]
3, [SU(3)C ]3,
[U(1)Y ]2SU(2)L, [U(1)Y ]
2SU(3)C , U(1)Y [SU(2)L]
2, U(1)Y [SU(3)C ]2, U(1)Y SU(2)LSU(3)C , [SU(2)L]
2SU(3)C
and SU(2)L [SU(3)C ]2. Starting with the Abelian anomaly, we have
[U(1)Y ]3 →
∑fL
tr [{gY YfL , gY YfL} gY YfY ]−∑fR
Tr [{gY YfR , gY YfR} gY YfR ] =
= 2g3Y
∑fL
Y 3fL −
∑fR
Y 3fR
= 2g3Y ng(6Y
3q + 2Y 3
l − 3Y 3u − 3Y 3
d − Y 3e ) = 0.
(2.99)
20
Next, using T i = σi
2 as the SU(2) generators, we obtain
[SU(2)L]3 → Tr
[{gT a, gT b
}gT c
]= g3 Tr
[1
2δabT c
]=g3
2δab Tr [T c] = 0, (2.100)
where the last equality follows from the fact that the T i are traceless. We can readily see that other pos-
sible anomalies, specifically [U(1)Y ]2SU(2)L, [U(1)Y ]
2SU(3)C , [SU(2)L]
2SU(3)C , SU(2)L [SU(3)C ]
3
and U(1)Y SU(2)LSU(3)C , vanish for the same reason (the SU(3) generators in the fundamental repre-
sentation are also traceless). Moving on to the pure SU(3)C anomaly we have
[SU(3)C ]3 → Tr
[{gSλa
2, gS
λb
2
}gSλc
2
]=g3S
2dabcnG
∑quarks
=g3S
2dabcng(2− 1− 1) = 0. (2.101)
There are only two anomalies left to investigate, U(1)Y [SU(2)L]2 and U(1)Y [SU(3)C ]
2:
U(1)Y [SU(2)L]2 → Tr
[{gT a, gT b
}gY Yf
]= g2gY Tr
[1
2δabI2Yf
]= g2gY nG
∑fL
Yf = 0 (2.102)
and
U(1)Y [SU(3)C ]2 → Tr
[{gSλa
2, gS
λb
2
}gY Yf
]= g2
SgY Tr
[1
3δabI3Yf + dabcT cYf
]=
= g2SgY ng
∑quarks
Yf = 0,(2.103)
where In is the identity matrix in n dimensions. We have shown that the SM is anomaly-free and it
should also be noted that this result does not depend on the number of generations (anomalies cancel
within each generation).
In the SM there is no attempt to describe gravitational interactions but, obviously, this can’t be the
case for a complete theory. Any possible inclusion of gravity in a gauge theory like this leads to new
anomalies related to local Lorentz transformations. In a four-dimensional Euclidean space these can
be considered SO(4) gauge transformations [36]. Due to the similarities between SO(4) and SU(2) we
conclude that only the mixed U(1)-gravity-gravity anomaly doesn’t automatically vanish [37]. Given that
all particles couple to gravity we have
U(1)-gravity-gravity→ Tr[{ggt
aSO(4), ggt
bSO(4)
}gY Yf
]= g2
ggY∑f
Yf = 0. (2.104)
Weather or not we consider a minimal extension to couple to gravity in four dimensions, the SM is an
anomaly-free theory.
21
Chapter 3
SU(5) based Grand Unified Theories
In this chapter, the SM’s most relevant problems and unattractive features are discussed and we will
see that Grand Unified Theories provide a favourable framework for attempts to solve some of these
issues. The GUTs that are treated in this thesis are based in an SU(5) gauge group so we study the
minimal SU(5) model as it can be considered a prototype for more complex theories introduced in later
chapters. After summarizing the minimal model’s flaws we will look at possible extensions that lead to
several improvements.
3.1 SM shortcomings and justification for GUTs
Despite its many successes and its historical importance, the SM has several insufficiencies when it
comes to describing interactions between fields in nature. Starting with a problem that has already
been mentioned in previous chapters, we note the absence of quantized gravitational interactions in
this model. In fact, this happens in all QFT as physicists have not, thus far, been able to gauge these
interactions. Until a breakthrough occurs in this area there is nothing that can be done to solve this
problem.
We move on to an issue where SM predictions are contradicted by experimental evidence: neutrino
masses. When the origin of fermion masses in the SM was explained, no Yukawa term for the neutrino
was introduced due to the absence of νR fields in the model. This was not a problem at the time the
theory was proposed because neutrinos were thought to be massless but nowadays neutrino oscillations
have been confirmed [38], which means that these particles must be massive. Even though there can
be no Dirac mass terms for the neutrinos, there is a type of mass term that only requires the existence
of one chirality state:1
2νTLC
−1MLνL +H.c., (3.1)
where C is the Dirac matrix for which Ψc ≡ CΨT and Ψc is the anti-particle of Ψ. This term is called
a Majorana mass, and it can only be gauge-invariant for fields that carry no conserved charge. Since
neutrinos are the only fermions with no electric charge it would seem like they could have Majorana
masses. They have non-zero lepton number, L, but non-perturbative effects (the so-called instanton
23
solutions) can violate individually the lepton number L and the baryon number B. Despite this, the
combination B−L remains invariant at the quantum level. This B−L symmetry prevents any dynamical
process that could generate a Majorana mass term, therefore neutrinos remain strictly massless in the
SM.
Another problem is the large number of arbitrary parameters. The scalar potential parameters are
constrained by viability and renormalization requirements while the Yukawa matrices are constrained
by experimental data but aside from that they can take a large number of values. The fact that so
many of the parameters of the theory have to be empirically observed is unsatisfactory, it would be
desirable to have a small number of physical parameters that must be measured and a series of physical
relations enabling the other parameters to be obtained from these. Furthermore, there are three fermion
generations in the SM but no explanation is provided for this, in particular, there is no a priori reason to
have the same number of lepton and quark doublets. This last aspect is vital for anomaly cancellation
so one could use this to justify the equality in the number of doublets but that is a contrived argument.
Theories with more constrained representations and Yukawa parameters provide a more favourable
framework to study the family structure and the repetition of the gauge representations. Aside from this,
it should be noted that the SM has no viable candidate for the dark matter that we know to exist in the
Universe.
A very unappealing feature of the SM is a so-called hierarchy problem related to the Higgs boson
mass. In (2.16) we did not consider radiative corrections, so this is just the ”bare” mass, not the one
measured in experiments. These corrections are of the order Λ2 where Λ is the cut-off scale. Data from
the LHC indicates that the Higgs mass is significantly lower than the cut-off scale, which means that the
radiative corrections must cancel each other almost completely. Put in another way, these corrections
must be fine-tuned in order to reproduce the empirical evidence. Although this is not necessarily a
problem, it is not natural and leads one to believe there must be some other way to get the correct Higgs
mass using physics beyond the SM.
We conclude this exposition of the SM’s shortcomings by noting that in this theory hypercharges are
assigned to the fields in order to get the correct electric charge, that is, the theory itself does not constrain
these quantum numbers. As a consequence, there is no electric charge quantization. On the other hand,
by requiring anomaly cancellation and making some additional reasonable assumptions it is possible to
obtain the hypercharges that give the correct electric charges as the only possibility [39, 40]. This,
however, is not a very satisfactory explanation, as in the case of the number of fermion doublets. More
promising solutions to this problem include the replacement of GSM with a larger group that contains the
former as a subgroup. The existence of this larger group would automatically lead to the quantization
of the electric charge and allow the correct charge of the elementary particles to be read from the
theory (instead of being put in by hand). In this scenario, since there is only one group determining the
interactions, there should only be one gauge coupling and the different couplings we observe at the SM
energy scale would be a consequence of symmetry breaking.
Theories in which there is gauge unification are called Grand Unified Theories (GUT) [16]. Before
proceeding with the discussion of GUTs we should check weather or not this idea of a unique coupling
24
at an energy scale above the SM one is feasible. This can be done since gauge couplings have an
evolution with the change in scale given by the Renormalization Group Equations (RGE). When the
GUT gauge symmetry is broken to GSM we have
αU = α1 = α2 = α3, (3.2)
where αU is the GUT gauge coupling and α1, α2 and α3 are the gauge couplings associated with the
GSM subgroups. At higher energies the GUT gauge symmetry is effective and the generators of these
subgroups obey a Lie algebra, therefore, certain relations between their normalizations must be verified:
αU = k1αy = k2αw = k3αs. (3.3)
The ki normalization factors are determined by the GUT gauge group being considered. Our interest is
in SU(5)-based theories, which belong to the canonical class [41], so ki ∝ (5/3, 1, 1).
The RGE of interest in this case is shown in Appendix A, here we only present the solutions to this
equation [31]:
α−1i (µ2) = α−1
i (µ1)− bi4π
ln
(µ2
2
µ21
), (3.4)
where the bi are one-loop beta coefficients that depend on group theory factors and the particle content
of a theory (a formula for calculating these coefficients is given in appendix A). Applying this equation to
the couplings being studied we get
α−11 (µ) = α−1
1 (MZ)− b12π
ln
(µ
MZ
), (3.5)
α−12 (µ) = α−1
2 (MZ)− b22π
ln
(µ
MZ
), (3.6)
α−13 (µ) = α−1
3 (MZ)− b32π
ln
(µ
MZ
). (3.7)
In order to take advantage of the well-known experimental values [42]
α−1(MZ) = 127.916± 0.015, (3.8)
αs(MZ) = 0.1184± 0.0007, (3.9)
sin2 θW = 0.23116± 0.00012, (3.10)
we may use (2.23) to write
α−11 (µ) = α−1(MZ)
3
5cos2 θW (MZ)− b1
2πln
(µ
MZ
), (3.11)
α−12 (µ) = α−1(MZ) sin2 θW (MZ)− b2
2πln
(µ
MZ
). (3.12)
25
Finally, we compute the SM’s beta coefficients, obtaining
b1 =41
10, b2 = −19
6, b3 = −7 (3.13)
and use these results in (3.5), (3.6) and (3.7) to make a plot of the SM gauge couplings dependence on
energy. This plot is shown in Fig. 3.1 and we can immediately observe that unification does not occur.
There is, however, no reason to give up on this goal because, as we have already seen, any theory
attempting to describe gaugeable interactions in a realistic way must include physics beyond the SM.
Also, adding to this, it can be stated that unification almost happens, so it shouldn’t be too difficult to
solve the problem in question when the particle content is enlarged as a consequence of new physics.
From Fig. and constraints related with proton decay (these will be discussed later) we can estimate the
unification scale to be between ∼ 6× 1014GeV and ∼ 1017GeV.
Figure 3.1: SM gauge couplings running.
As we want to test the possibility of unification in other contexts it is useful to introduce a tool that
can facilitate this analysis, the B-test [43]. Taking into account contributions to the beta coefficients from
new fields with masses between the electroweak and unification scales we can define an effective beta
coefficient [44]
Bi ≡ bi +∑I
bIi rI , (3.14)
where the bIi represent the contribution from particle I with mass given by MI and the rI are ratios that
determine how relevant a particle’s contribution is as a function of its mass, specifically
rI =ln(Λ/MI)
ln(Λ/MZ). (3.15)
26
Defining Bij = Bi −Bj we obtain the following B-test,
B ≡ B23
B12=
sin2 θW − αα3
35 −
85 sin2 θW
, (3.16)
as well as the GUT scale relation
B12 ln
(Λ
MZ
)=
2π
5α(3− 8 sin2 θW ). (3.17)
The equation in (3.16) makes it easier to test unification because, while the left-hand side depends
on the particle content of the theory, the right-hand side depends only on group theory factors and
quantities that are measured at low energies, so its value is fixed once we pick a particular class of GUT
to work with. We conclude that, in an SU(5) framework, unification requires
B = 0.718± 0.003, (3.18)
while in the SM we have B ' 0.53, which means that we have to increase the value of B.
3.2 Minimal SU(5)
We begin this section by explaining the choice of SU(5). To meet the criteria of providing a GUT, a
gauge group must be simple or a direct product of identical simple groups so that the gauge coupling
is unique. On the other hand, this group should coincide with GSM at low energies, which can happen
only if it embeds the SM group. The couplings are unified when the larger (GUT) group is effective but
as the energy scale gets lower spontaneous symmetry breaking occurs and the remaining gauge group
will be GSM . Since this group has rank 4 (4 generators that can be simultaneously diagonalized), any
GUT gauge group must have rank 4 or higher. There are many possibilities for which these conditions
are met but we are not looking for a radically different theory; the best course of action is to consider the
simplest candidates.
In this context, the SU(5) gauge group constitutes a favourable choice. It contains GSM and has the
same rank as it, while also having a minimal particle content, that is, among all possible GUTs based on
rank 4 groups it is the one in which the number of fields that have to be added to the SM particle content
is smallest. In this section we will study the minimal SU(5) model, which, as the name indicates, is a
minimal extension of the SM into a GUT. As will later be seen, this model has many flaws and is clearly
not correct but it serves as a basis on which more sophisticated SU(5) GUTs can be built.
According to what was already said, there is a spontaneous symmetry breaking form SU(5) to GSM
and from this last group to SU(3)C × U(1)Q. For reasons that will be explained later, the new gauge
bosons implied by the larger gauge group must have masses several orders of magnitude larger than
the SM gauge bosons. In order to have two different mass scales in the theory we need more scalar
fields than those present in the SM. In the context of SU(5) GUTs this can be achieved using two scalar
representations that acquire a vev, a 24H adjoint representation that breaks SU(5) and a 5H fundamental
27
representation that contains the SM Higgs doublet. The adjoint representation is chosen because, as we
will see further ahead, there is a minimum of the scalar potential associated with this representation for
which the rank of the group is preserved when symmetry break occurs (otherwise it would be broken, a
clear problem since SU(5) and GSM have the same rank) and SU(5) is broken down to GSM [31]. As for
the 5H , the fact that GSM is a maximal subgroup of SU(5) means that the fundamental representation of
SU(5) can be constructed using the fundamental representations of SU(3) and SU(2): 5 = (3, 1)⊕ (1, 2).
The SU(2) fundamental representation has the right quantum numbers for the Higgs doublet but, now,
there is an additional scalar colour triplet corresponding to the fundamental representation of SU(3).
This triplet has some important consequences that will be discussed ahead.
When SU(5) undergoes spontaneous symmetry breaking, there is more than one group it can break
to. In order for the breaking to occur in the desired direction, the scalar potential associated with 24H
must verify certain conditions. The most general renormalizable (of order 4 in the scalar fields) and
SU(5) invariant scalar potential has the form
V = V (24H) + V (5H) + V (24H , 5H), (3.19)
with
V (24H) = −µ224
2Tr{242
H}+λ2
4Tr{242
H}2 +λ3
4Tr{244
H}+a1
3Tr{243
H}, (3.20)
V (5H) = −µ25
25†H5H +
λ1
4(5†H5H)2 (3.21)
and
V (24H , 5H) = λ185†H5HTr{
242H
}+ λ195†H242
H5H + a35†H24H5H , (3.22)
where µ24, λ1, λ2, λ3, µ5, a1, λ18, λ19 and a3 are constants. The notation used for these constants is
related to the fact that we will work with more scalar representations in later chapters (the most general
renormalizable and SU(5)-invariant potential in that context is shown in Appendix B).
Similarly to what happens in the SM, there is spontaneous symmetry breaking if V (24H) has a
non-zero minimum. It can be shown that (3.20) has three possible extrema [31], σdiag(2, 2, 2,−3,−3),
v41diag(1, 1, 1, 1,−4) and diag(0, 0, 0, 0, 0), where σ and v41 are real constants related to λ2, λ3, a1 and
µ24, and the first of these extrema is the one associated with breaking into GSM . To see how this ex-
tremum is obtained we start by noting that 24H can be diagonalized through a unitary transformation:(24H)ij →
(24H)iδij , with
∑i
(24H)i = 0. In this context the equations ∂V (24H)/∂Hi = 0 are cubic equations in the
diagonal elements, Hi, which can, then, assume at most three different values. Detailed calculations
(which are beyond the scope of this work) show that we obtain the desired minimum when λ3 > 0 and
λ2 > −7/3λ3 [31], with
σ2 =µ2
24
(30λ2 + 7λ3)(3.23)
and in arriving at this last equation we have, for simplicity, imposed extra Z2 discrete symmetries 24H →
−24H , 5H → −5H in order to get rid of cubic terms.
An SU(n) group has n2 − 1 generators, therefore, SU(5) has 24 generators. These are usually
28
represented by the generalized Gell-Mann matrices, that can be found in Appendix B. Observing those
matrices one notices that the breaking of SU(5) to GSM occurs in the λ24 direction. In fact, as we will
see further ahead, this generator’s eigenvalues are the SM hypercharges.
After the SU(5) gauge breaking some fields associated with 24H acquire a mass. To see how this
happens we start by decomposing the adjoint representation in terms of GSM quantum numbers, ob-
taining
24H = Σ8 ⊕ Σ3 ⊕ Σ(3,2) ⊕ Σ(3∗,2) ⊕ Σ0, (3.24)
where Σ8 is an SU(3) octet, Σ3 is an SU(2) triplet and Σ0 is a singlet (the other 2 field’s notation is
self-explanatory, SU(3) and SU(2) quantum numbers for all these fields are shown explicitly in Appendix
B). Since it is the singlet that has a vev responsible for breaking SU(5), we proceed in a similar way to
what was done in the SM by shifting the field to obtain a new set of scalars, which may be expressed as:
24′H = 24H − 〈24H〉 =
Σ8 − 2Σ0/√
30 Σ(3,2)
Σ(3∗,2) Σ3 + 3Σ0/√
30
. (3.25)
The masses of these fields can be obtained by evaluating the second derivative of the scalar potential
at H = 〈H〉, yielding [31]
m2Σ8
= 5σ2λ3, m2Σ3
= 20σ2λ3, m2Σ0
= 2µ224, m
2Σ(3,2)
= m2Σ(3∗,2)
= 0. (3.26)
Through a Higgs Mechanism the massless fields in this last equation are absorbed by the theory’s new
gauge bosons as longitudinal degrees of freedom and these latter become massive.
The 5H will acquire a vev at the SM scale, which is much lower than the GUT scale, so when the
adjoint scalar is in its vev state the effective scalar potential is given by
Veff(5H) = −µ25
25†H5H +
λ1
4(5†H5H)2 + λ185†H5HTr{〈24H〉2}+ λ195†H〈24H〉25H + a35†H〈24H〉5H . (3.27)
By rearranging the terms, recalling the previously mentioned extra Z2 discrete symmetry and by sepa-
rating the 5H into the SM doublet, H, and the colour triplet, T, we get
H†H(−µ25
2+σ2
15(30λ18 + 9λ19)) + T †T (−µ
25
2+σ2
15(30λ18 + 4λ19)) +
λ1
4(5†H5H)2. (3.28)
From this expression, we conclude that 5H fields have masses given by
m2T =
µ25
2− (30λ18 + 4λ19)σ2, (3.29)
m2H =
µ25
2− (30λ18 + 9λ19)σ2. (3.30)
We already saw that the scalar doublet H plays the role of the SM Higgs, so its mass is expected to be
very low when compared to σ2 in order for it to survive at low energies while the heavy particles with
masses close to σ decouple. Recalling (3.21), we note that this is equivalent to the scalar potential in
29
the SM, so H has a vev of
〈H〉 =1√2
0
v
(3.31)
with v = (4m2H/λ1)
12 .
Since the mass of H is expected to be of the order of the electroweak scale while mT should be
many orders of magnitude larger for reasons that will be made clear in the next chapter, we have a
problem, known as the doublet-triplet splitting problem: how can two scalars that belong to the same
representation have so different masses? This is not a formal problem, nothing forbids this a priori, but
it doesn’t seem natural. Looking at this in another way, one can show that, with the previously shown
scalar potential, the W boson mass will be
M2W =
1
4g2v2 =
g2
2λ1(µ2
5 + 15σ2(−4λ18 +6
5|λ19|)). (3.32)
For this formula to be consistent with the experimental value of MW (of the order of the SM scale) the
parameters λ18 and λ19 must be fine-tuned.
There are 24 gauge bosons in the minimal SU(5) model, because the gauge group has 24 genera-
tors. The covariant derivative for the fundamental transformation is, then, given by
Dµ = ∂µ + ig5
23∑a=0
Aaµλa
2= ∂µ + ig5Aµ, (3.33)
where g5 is the gauge coupling and λa are the 24 generalized Gell-Mann matrices that represent the
SU(5) generators (in the fundamental representation). These 24 gauge bosons include those present in
the SM and 12 new ones:
G11µ +
2Bµ√30
G12µ G1
3µ X1cµ Y 1c
µ
G21µ G2
2µ +2Bµ√
30G2
3µ X2cµ Y 2c
µ
G31µ G3
2µ G33µ +
2Bµ√30
X3cµ Y 3c
µ
X1µ X2
µ X3µ
Zµ√2−√
310Bµ W+
µ
Y 1µ Y 2
µ Y 3µ W−µ −Zµ√
2−√
310Bµ
. (3.34)
The new X and Y bosons constitute SU(3) triplets and SU(2) doublets, therefore, they can connect a
fermion line with a quark line (for this reason they are sometimes called leptoquarks). This means that
there are new interactions not present in the SM, some of which may lead to proton decays. The colour
triplet T that is included in 5H contributes to these processes as well. An analysis of proton decays in a
GUT model is deferred to the next chapter, for now it is sufficient to know that experimental data related
with this phenomenon imposes constraints on the masses of particles responsible for it.
Spontaneous symmetry breaking of the SU(5) leads to the generation of mass for the 12 gauge
bosons that are not present in the SM through the Higgs Mechanism. All the generators associated with
these bosons are broken, since the surviving group will be GSM . By computing the kinetic terms of the
24H , the terms that include covariant derivatives of these fields, it can be seen that the X and Y bosons
30
acquire a mass
M2X = M2
Y =25
2g2
5σ2, (3.35)
where σ is the vev of the 24H .
The SM and minimal SU(5) theories have the same fermion fields. In the latter they are distributed
between an antifundamental 5 representation and an antisymmetric 10 representation (for each family):
5F =
dc1
dc2
dc3
e−
−νe
, (3.36)
10F =
0 uc3 −uc2 u1 d1
−uc3 0 uc1 u2 d2
uc2 −uc1 0 u3 d3
−u1 −u2 −u3 0 e+
−d1 −d2 −d3 −e+ 0
(3.37)
This assignment yields the correct quantum numbers and solves one of the SM’s problems: it is, now,
possible to read the hypercharge from the λ24 generator of SU(5); the hypercharge and, consequently,
the electric charge, are no longer put in by hand to reproduce experimental results, they are a conse-
quence of the representation the particle belongs to and the eigenvalues of the hypercharge operator.
This also provides an explanation for the fractional charge of the quarks.
Fermion masses arise in the same way they did in the SM. The only SU(5) and Lorentz invariant
terms that can be constructed using the fermion and scalar fields of the theory are
LY =1
8ε510TFCY1010F 5H + 5TFCY510F 5∗H +H.c., (3.38)
where the Y’s are the arbitrary Yukawa matrices. The mass terms will be the ones involving the Higgs
doublet, H. When these terms are expanded one obtains the following relationships:
Ye = Y Td (3.39)
and
Yu = Y Tu . (3.40)
We can see that, in this model, the Yukawa matrices are not completely arbitrary and there are only
2 of them, in contrast with the 3 in the SM. Additionally, (3.39) means that the down quarks have the
same masses as the charged leptons at the GUT scale (when the GUT group is effective). To check if
this may be viable one can run these masses from the SM scale to the unification scale using the RGE
and reasonable estimates for the unified scale and coupling. Even when the large errors are taken into
31
account it is clear that this last result is wrong. The minimal SU(5) model is necessarily incorrect.
Like in the SM, neutrino masses are absent in the minimal SU(5) model. The fact that there are no
right handed neutrinos in the theory means that there cannot be tree-level renormalizable Dirac mass
terms for the neutrino. In this scenario light neutrino masses can still be generated through the so-called
seesaw mechanisms but these are not included in the minimal model. Extensions of the SM or the
theory studied in this section may have such mechanisms but in the SM they would be very heavy and
unstable. GUTs, on the other hand, provide a new (heavy) scale of physics and it is conceivable that
there can be some SU(5) field inducing a seesaw mechanism. The explanation of these phenomena will
be given when extensions to the minimal model are discussed.
Xµ Σ8 Σ3 T
B23 − 113 rX -rΣ8
23rΣ3 − 1
6rT
B12 − 223 rX 0 − 2
3rΣ3
115rT
Table 3.1: New minimal SU(5) fields that contribute to the B-test and their influence.
We have been referring to the minimal SU(5) as a GUT but it is still necessary to see if unification
may be achieved in this model. The particles that are absent in the SM and contribute to gauge coupling
running are the leptoquarks, T, Σ8 and Σ3. Using data from Appendix A we are able to compute these
field’s contributions to the value of B ≡ B23
B12. The results are shown in Table 3.1 and, since we want
to increase B, it is clear that the leptoquarks and Σ3 favour unification while T and Σ8 worsen it. For
the theory to be consistent with proton decay empirical data, leptoquark boson masses are expected
to be around unification scale order, so rXµ is very small and contributions from these fields can be
neglected. Looking at (3.26) we can see that, unless λ3 has an unnaturally small value, mΣ3won’t
be far from σ and contributions from this field will not be relevant. Furthermore, since mΣ8 has the
same order of magnitude as mΣ3and Σ8 is unfavourable for unification, the former mass would have to
be unrealistically small for this goal to be reached. We conclude that the minimal SU(5) model is not
actually a GUT, which is a significant problem but, as already stated, this model serves as a framework
for the study of general GUT features and as a basis for the construction of realistic theories so it still has
interest and it is possible that certain extensions eliminate this issue. As we show in the next section,
frequently used means of fixing the mass mismatch or adding neutrino masses involve extending the
particle content so it may be possible to unify gauge couplings in these SU(5)-based models.
32
3.3 SU(5) Extensions
As we saw in the previous section, the minimal SU(5) has several problems that make it incompatible
with experimental data. Despite this, the fact that this model quantizes electric charge and predicts the
correct charges for all fermions, combined with the fact that unification almost happens, giving estimates
for the unification scale that are consistent with the bounds imposed by proton decay, justify further
pursuit of SU(5) GUTs. The rest of this chapter will be a summary of some frequently seen extensions
of the minimal model. These extended models will be divided in two classes, according to the way
in which the problem of the wrong relationship between the masses of the down quarks and charged
leptons is solved: there are non-renormalizable models, where the particle content doesn’t need to be
enlarged and renormalizable models, where the particle content has to be enlarged.
3.3.1 Non-Renormalizable Models
In this case the Yukawa sector is changed by addition of non-renormalizable terms. A theory like this
is, by definition, incomplete. The reason why it is still worthwhile to work is such a set up is that, since
there are other problems anyway, one can consider the existence of a larger theory that will ultimately
solve all the problems and give consistent effective low energy Yukawas. The non-renormalizable terms
added are [45]
∆LY =5FY(1)5 10F
(Φ
Λ5H
)∗+ 5FY
(2)5
(Φ
Λ10F
)5∗H +
1
8ε510FY
(1)10 10F
(Φ
Λ5H
)+
+1
8ε510FY
(2)10
(Φ
Λ10F
)5H +H.c.,
(3.41)
where Φ is a scalar singlet and Λ is the cut-off scale where the effective operators stop being valid. It
can be taken to be the Planck scale. Considering only the terms with the Higgs doublet we get
Yd = Y T5 −√
3
5
v
ΛY
(1)5 +
2√15
v
ΛY
(2)5 (3.42)
Ye = Y5 −√
3
5
v
ΛY
(1)5 −
√3
5
v
ΛY
(2)5 (3.43)
Yu = −1
2(Y10 + Y T10) +
3
2√
15
v
Λ(Y
(1)10 + Y
(1)T10 )− 1
4√
15
v
Λ(2Y
(2)10 − Y10(2)T ). (3.44)
These new Yukawa matrices have enough adjustable parameters to fit the experimental data.
There is still no gauge unification because the new Yukawas do not alter the running of the theory’s
gauge couplings. This unification can be achieved if new fields are added. At the same time, one
can add certain fields to generate light neutrino masses through seesaw mechanisms. Therefore, both
the unification and the neutrino mass problems can be simultaneously solved by introducing the fields
necessary for the occurrence of seesaw mechanisms that will be described in the end of this chapter.
33
3.3.2 Renormalizable Models
It is possible to ”fix” the relationship between down quark and charged lepton masses of the minimal
model without sacrificing renormalizability. This is done by adding a new scalar representation that can
couple with the fermion fields without changing the breaking pattern. In order to achieve this, we choose
a representation that acquires a non-vanishing vev alongside the SM Higgs by participating in the Higgs
Mechanism. For this to happen the new representation should contain an SU(2) doublet and have an
SU(3)C × U(1)Q invariant vev structure.
A convenient and usual choice is the 45H [46], which verifies the previously stated conditions. It is
derived from 10⊗ 5 = 5⊕ 45 and can be decomposed as
45 = (8, 2, 1/2)⊕ (6, 1,−1/3)⊕ (3, 2,−7/6)⊕ (3, 1, 4/3)⊕ (3, 3,−1/3)⊕ (3, 1,−1/3)⊕ (1, 2, 1/2)), (3.45)
in terms of GSM quantum numbers. The penultimate representation can be identified as a new scalar
colour triplet, T2, and the last one can be identified as a new SU(2)doublet scalar, H2. We can write this
decomposition in a more convenient notation as
45H = S(8,2) ⊕ S(6∗,2) ⊕ S(3∗,2) ⊕ S(3∗,1) ⊕∆⊕ T2 ⊕H2. (3.46)
The vev of the 5H and 45H fields need to be in a correct configuration if we want to reproduce the
SM spontaneous symmetry breaking. For the 5H this is given by
〈5H〉T = (0, 0, 0, 0, v5). (3.47)
In the 45H case, we note that this representation obeys
45αβγ = −45βαγ (3.48)
because the upper indexes come from the completely anti-symmetric 10 representation, which also
implies
(45H)ααγ = 0. (3.49)
Considering this, the vev structure is [47,48]
〈45α5Hβ〉 = v45(δβα − 4δα4 δ
4β), α, β = 1, ..., 4. (3.50)
Finally, these vev combine to give the SM’s v through [49]
v2 ≡ |v5|2 + 24|v45|2 = (√
2GF )−1, (3.51)
where GF is the Fermi constant.
With the addition of the new scalar representation the most general Yukawa sector of the Lagrangian
34
is, in simplified form
− LY =ε54
((Γ1u)ij10i10j5H + (Γ2
u)ij10i10j45H) +√
2((Γ1d)ij10i5
∗j5H + (Γ2
d)ij10i5∗j45∗H), (3.52)
where i,j are generation indices and the Γs are Yukawa matrices. It should be noted that Γ1u is symmetric
and Γ2u is anti-symmetric. Substituting (3.47) and (3.50) in (3.52) we find that, when 5H and 45H acquire
vevs, we get new mass matrices for the fermions:
Mu = v′Γ1u + 2v45Γ2
u, (3.53)
Md = v′∗Γ1d + 2v∗45Γ2
d, (3.54)
MTe = v′∗Γ1
d − 6v∗45Γ2d. (3.55)
A consequence of this is that
Md −MTe = 8v∗45Γ2
d, (3.56)
so the charged lepton and down quark mass matrices are no longer predicted to be equal above unifi-
cation scale. Notice that the presence of Γ1d in Md and Me is very important for these two last matrices
to be different in a way that can solve the mass mismatch problem, otherwise they would just be differ
by a multiplicative factor and the problem would remain.
S(8,2) S(6∗,1) S(3∗,2) S(3∗,1) ∆ T2 H2
B23 − 23rS(8,2)
− 56rS(6∗,1)
16rS(3∗,2)
− 16rS(3∗,1)
23r∆ − 1
6rT2
16rH2
B12 − 815rS(8,2)
215rS(6∗,1)
1715rS(3∗,2)
1615rS(3∗,1)
− 95r∆
115rT2
− 115rH2
Table 3.2: Contributions to the B-test from 45H fields.
The new fields associated with 45H alter the running of SM gauge couplings, with their contributions
to the B test displayed in Table 3.2. Most of these fields have an unfavourable impact on unification
attempts, with the exceptions being ∆ and H2. This last field has a very positive effect as its mass is
expected to be low (of the electroweak scale order). As for ∆, this field contributes to proton decay so
its mass is constrained to be around 1010 GeV or higher. However, if the fields from 45H that make B
smaller have high masses that turn their contributions negligible, this value for the ∆ mass is consistent
with unification.
Extending the minimal SU(5) model by adding a 45H scalar representation leads to a theory where
gauge coupling unification may occur and the mass mismatch problem is no longer present. This comes
at the cost of increasing the theory’s number of parameters, while new naturality problems arise as
35
proton decay and unification constraints force fields from the same 45H representation to have very
different masses but, overall, this extension improves the model, making it more realistic and a proper
GUT. Although there are still no neutrino masses this issue can be solved, as we will see, by introducing
seesaw mechanisms.
3.3.3 Seesaw Mechanisms
As pointed out in the beginning of this chapter, a very relevant deficiency of the SM is the absence of
neutrino masses, but this problem is yet to be addressed in the context of GUTs. Although we could
consider simply adding right-handed neutrinos to the theory so these particles would acquire mass in the
same way as all other fermions, the relative smallness of neutrino masses suggests a different approach:
seesaw mechanisms. In the rest of this chapter we will make a brief review of the most usual types of
seesaw (types I, II and III) and this will be done having the SM as background because it is simpler to
understand these processes in this framework and because our main interest regarding neutrinos is in
their low-energy phenomenology (that is, at energy scales bellow the GUT one, where the gauge group
is GSM ). Furthermore, as we will see in the next chapter, once these mechanisms are understood they
can easily be applied in the context of GUTs.
Figure 3.2: Diagram associated with th d=5 Weinberg operator.
Seesaw mechanisms are characterized by the introduction of new heavy particles in the theory,
with masses that may reach the GUT scale. Tree-level exchange of these particles (depicted in Fig.
3.2) generates a d=5 Weinberg operator, which is the lowest order operator that can lead to Majorana
neutrino masses [50, 51]. The Lagrangian term associated with this is, after integrating-out the heavy
fields,
LWeinberg = −zαβ
Λ(lLαH)C(lLβH)T +H.c., (3.57)
where Λ is the high-energy physics cutoff scale and zαβ are complex constants. Upon SSB this last
equation becomes
LWeinberg = −1
2mαβν ¯νLαν
cLβ +H.c.+ ..., (3.58)
where mαβν = v2zαβ/Λ is the 3 × 3 effective neutrino matrix. It can be seen in what comes next that
36
larger masses of the new fields lead to smaller mνs, hence the designation of seesaw. It should be
pointed out that the calculations performed until the end of this chapter are not exact as only first-order
perturbation theory is used and some channels are ignored in the diagrams that are studied but our
simplified procedure is enough, in this high mass limit, to clarify the flavour structure of the neutrino
mass matrix and its dependence on the new field’s masses.
Type 1 Seesaw
In this type of seesaw the new heavy particles are nR right-handed neutrino fields νRi, with quantum
numbers given by (1,1,0) [52, 53]. Taking this into account, the Lagrangian of the original theory is
modified to
LI = LSM +i
2νRi�∂νRi − Y αiν lLαHνRi −
1
2M ijR
¯νcRiνRj +H.c., (3.59)
where Yν is a 3 × nR complex and arbitrary Yukawa matrix while MR is a nR × nR symmetric matrix.
Before seeing how the exchange of these νRi leads to small neutrino masses it is convenient to change
basis in order to have the new fields as mass eigenstates:
νRi = RijRNRj , RTRMRRR = dR = diag(M1, ...,MnR). (3.60)
Figure 3.3: Feynman diagram representing the exchange of heavy particles that generates type I see-saw.
The Lagrangian in (3.59) may now be written as
LI = LSM + iNRi�∂NRi − Y αiR lLαHNRi −1
2dijRN
cRiNRj +H.c., (3.61)
where YR = YνRR. Comparing the Feynman diagram for νRi exchange (Fig. 3.3) with Fig. 3.2 and
comparing (3.61) with (3.57) we obtain
zαβ
∆∝ Y αiR
1
�p−MiY βiR . (3.62)
It has already been stated that the Mi are much larger than the electroweak scale so we can make the
following approximation:zαβ
∆' −Y αiR
1
MiY βiR = −Y αiR
1
diiRY βiR . (3.63)
37
Recalling (3.60) we can write
zαβ
∆'− Y αiR (diiR)−1Y βiR = −(YRd
−1R Y TR )αβ = −(YνRR(RR)−1(MR)−1(RTR)−1RTRY
Tν )αβ =
= −(YνM−1R Y Tν )αβ .
(3.64)
To get the desired result, the effective mass matrix of light neutrinos, we use (3.57):
mν = −v2YνM−1R Y Tν = −mDM
−1R mT
D, (3.65)
with mD = vYν . We can easily see that large values of MR lead to small mν , as we wanted in order to
reproduce experimental data.
Type 2 Seesaw
In type II seesaw we introduce (ar least) one SU(2) scalar triplet ∆ = {∆1, ∆2, ∆3} with quantum num-
bers given by (1,3,1) [54–58]. As these fields are not singlets, some care must be taken to ensure that
terms involving them are gauge invariant. We can use the Pauli matrices to write them as
∆ = ∆iσi
2=
1
2
∆3 ∆1 − i∆2
∆1 + i∆2 −∆3
. (3.66)
Choosing the basis where T 3 is a 3× 3 diagonal matrix, the following relations hold:
∆ =
∆+√
2−∆++
∆0 −∆+√
2
⇒ Q(∆) =
∆+√
2−2∆++
0 −∆+√
2
. (3.67)
As the new scalar fields transform under SU(2) according to
´∆ = e−iT
aωa∆, (T a)ij = −iεaij , (3.68)
we see that ∆ transforms according to ´∆ = U∆U†, where U is an unitary matrix. The possible terms
involving ∆, namely lL∆†lcL and HT ∆∆ are still not gauge invariant as
¯lL∆†C
¯lTL = lLU
†U∆†U†U∗ClLT
= lL∆†U†U∗ClLT, ´HT ´∆ ´H = HTUTU∆U†UH = HTUTU∆H.
(3.69)
Finally, to achieve invariance, we rotate the fields using the Pauli matrix σ2, as we did for the Higgs field
in the previous chapter:
∆ = iσ2∆ =
∆0 −∆+√
2
−∆+√
2∆++
. (3.70)
The fields ∆ transform as ∆ = U∗∆U† and the possible terms associated with the new scalars are
38
Figure 3.4: Feynman diagram representing the exchange of heavy particles that generates type II see-saw.
now invariant, so the Lagrangian (in first order approximation) becomes
LII = LSM + Tr[(Dµ∆)†(Dµ∆)
]+M2
∆ Tr[∆†∆
]− (Y αβ∆
¯lLα∆†lcLβ − µHT∆H +H.c.) + ..., (3.71)
where Y∆ is a 3× 3 symmetric matrix, M∆ is the mass of ∆ and µ is a coupling constant. Interactions of
∆ exchange are shown if Fig. 3.4 and, following the same procedure that was used for Type I seesaw
we arrive atzαβ
Λ∝ Y αβ∆
1
p2 −M2∆
(−µ) ' λ
M∆Y αβ∆ , (3.72)
where λ = µM∆
is an adimensional parameter. From this we get (after SSB)
mν =vλ
M∆Y∆. (3.73)
Generalization to the case in which multiple scalar triplets are added is simple and yields
mν =
n∆∑i=1
vλiM∆i
Y∆i, λi =
µiM∆i
. (3.74)
Type 3 Seesaw
Type III seesaw is obtained by adding nΣ SU(2) fermion triplets ΣRi, with quantum numbers given by
(1,3,0), to the theory [59]. Before showing the changes to the Lagrangian it is convenient to rewrite these
new fields in a way that facilitates their inclusion in gauge invariant terms. Using the definition of electric
charge and the Pauli matrices we get
ΣRi =
Σ0Ri√2−Σ+
Ri
Σ−Ri −Σ0Ri√2
. (3.75)
To have the Lagrangian in the same form as in the previous cases we can redefine (3.75) as
ΣRi =
Σ0Ri
√2Σ+
Ri√2Σ−Ri −Σ0
Ri
(3.76)
39
and, finally,
LIII = LSM +i
2tr(ΣRiDΣRi)− Y αiT lLαΣRiH −
1
2M ij
Σ tr(ΣRiΣRj) +H.c., (3.77)
where YT is a 3 × nΣ arbitrary complex Yukawa matrix and MΣ is a nΣ × nΣ symmetric matrix. In this
case the Weinberg operator is generated by the exchange process shown in Fig. 3.5 and, due to the
similarities between this situation and Type I seesaw we can obtain the light neutrinos effective mass
matrix simply by comparing (3.77) with (3.61):
zαβ
Λ' −(YTM
−1Σ Y TT )αβ ⇒ −v2YTM
−1Σ Y TT = −mTM
−1Σ mT
T , (3.78)
where mT = vYT .
Figure 3.5: Feynman diagram representing the exchange of heavy particles that generates type IIIseesaw.
40
Chapter 4
Discrete Symmetries and Proton
Decay
In this chapter we introduce the Adjoint SU(5) model with intention of building a realistic and predictive
theory from it. We investigate proton decay and the possibility of using discrete symmetries to simulta-
neously alleviate constraints related to this process and reduce the number of Yukawa parameters.
4.1 Adjoint-SU(5) Model
The Adjoint SU(5) model can be considered a natural consequence of our discussion so far because it
is obtained from the minimal SU(5) model by adding a 45H scalar representation to solve the mass mis-
match problem and nρ adjoint fermionic fields ρ(24) to generate light neutrino masses through seesaw
mechanisms [17]. This choice of the adjoint representation follows from the fact that, decomposing it in
terms of GSM quantum numbers,
ρ(24) = ρ8 ⊕ ρ3 ⊕ ρ(3,2) ⊕ ρ(3∗,2) ⊕ ρ0, (4.1)
one sees that it contains a singlet and an SU(2) triplet so both Type I and Type III seesaw are induced.
The main effects of adding a 45H scalar to the minimal SU(5) were already covered in section 3.3.2,
therefore, we will now focus on changes brought on by the ρ(24) fields. We assume that these fields
acquire unconstrained Majorana masses, so the most general Yukawa sector consistent with renormal-
izability is now given by
−LY =ε54
((Γ1u)ij10i10j5H + (Γ2
u)ij10i10j45H) +√
2((Γ1d)ij10i5
∗j5H + (Γ2
d)ij10i5∗j45∗H)+
+MklTr(ρkρl) + λklTr(ρkρlΣ) + (Γ1ν)ik5∗i ρk5H + (Γ2
ν)ik5∗i ρk45H +H.c.,
(4.2)
where k, l = 1, ..., nρ, M is a symmetric nρ × nρ Majorana mass matrix and λ, Γ1ν and Γ2
ν are Yukawa
matrices (λ is symmetric).
41
When the adjoint scalar 24H gets a vev we have ρ(24) masses given by
M0 =1
4(M − σ√
30λ),
M3 =1
4(M − 3σ√
30λ),
M8 =1
4(M +
2σ√30λ),
M(3,2) = M(3∗,2) =1
4(M − σ
2√
30λ).
(4.3)
From now on, since we are interested in analysing Types I and III seesaw, we will no longer consider the
ρ8, ρ(3,2) and ρ(3∗,2) fields. Looking at (3.52) and the analogous equations, (3.56) and (3.72) we notice
that the light neutrino masses are proportional to v2, where v is the SM Higgs doublet’s vev. On the
other hand, we saw in section 3.3.2. that, in a framework with a 45H scalar representation, this constant
v appears as a combination of the H1 and H2 vevs. It is, then, convenient to make a basis change in
order to have one Higgs double H with a vev of v and another one, H, without vev:HH
=
cosβ sinβ
− sinβ cosβ
H1
H2
, (4.4)
with tanβ ≡ v45/v5. In this new basis we get, from (4.2) and the decompositions of ρ(24) fields shown
in Appendix B, the following Yukawa interactions involving these fields and SM fermions:
−LY =
√15
2
[cosβ
5(Γ1ν)ik + sinβ(Γ2
ν)ik
]lTi iσ2ρ0kH+
+
√15
2
[− sinβ
5(Γ1ν)ik + cosβ(Γ2
ν)ik
]lTi iσ2ρ0kH+
+1√2
[− cosβ(Γ1
ν)ik + 3 sinβ(Γ2ν)ik
]lTi iσ2ρ3kH+
+1√2
[sinβ(Γ1
ν)ik + 3 cosβ(Γ2ν)ik
]lTi iσ2ρ3kH +H.c..
(4.5)
Comparing this result with the descriptions of Type I and Type II seesaw in the last chapter, we obtain
(mν)ij = −(mD0 M
−10 mD
0
T)ij − (mD
3 M−13 mD
3
T)ij , (4.6)
with
mD0 =
√15v
2(cosβ
5Γ1ν + sinβΓ2
ν), mD3 =
v√2
(− cosβΓ1ν + 3 sinβΓ2
ν). (4.7)
Since M0 and M3 are unconstrained Majorana masses they can have very high values, making the left-
handed neutrino masses very small, in accordance with experimental evidence. There is another type
of seesaw that can be induced by the ρ8, radiative seesaw [60–63], but it was ignored for simplicity sake.
We can do this because the new contributions are very suppressed if one assumes mρ3k � mρ0k and
mρ8k � mρ3k [64,65].
A discussion of unification and associated constraints is postponed until the next chapter, where we
42
choose a particular value of nρ and a discrete symmetry is incorporated in the theory.
4.2 Proton Decay in the Adjoint SU(5) model
It has been mentioned, in the previous chapter, that proton decay is a common problem in GUTs. This
process is absent in the SM due to an accidental symmetry that preserves B − L and the fact that the
proton is the lightest baryon. In GUTs this symmetry is lost as there are new particles that can mediate
proton decay, which constitutes a problem for such theories because no event of this kind has been
observed so far. This last aspect means that protons have a very large lifetime and, consequently, any
amplitude that contributes to the process must be very small if a model is to describe elementary particle
interactions correctly.
The sources of proton decay in the Adjoint model are the leptoquarks and coloured scalar triplets.
When theX and Y bosons were introduced it was pointed out that they can connect quark lines to lepton
lines, since they are SU(3) triplets and SU(2) doublets. Because of this there will be new interactions
that are forbidden in the SM and it is important to study them in order to discover their possible conse-
quences. The interactions between bosons and fermions come from the terms with covariant derivatives
of the fermion fields, which are given by
LF = i1
2Tr{Ψ10��DΨ10}+ iΨ5��DΨ5, (4.8)
so one gets
LFint = −g5Tr{Ψ10γµAµΨ10}+ g5Ψ5γ
µATµΨ5. (4.9)
We are only interested in the new interactions so the terms of relevance in this context are the ones
involving the new bosons:
LXY = −g5Tr{Ψ10γµAXµ Ψ10}+ g5Ψ5γµ(AXµ )TΨ5. (4.10)
Since the X and Y bosons belong to the same SU(5) representations and have the same quantum
numbers, as well as the same mass, they can be regarded as the same field with two indexes, that is,
(Xµ)aα where a is an SU(3)C index, α is an SU(2)L index and a = 1→ X, a = 2→ Y . With this in mind,
(4.10) may be written as
LXY =g5√
2
[((dc)
αγµεabLb − ecεbaγµqαb + qβaγ
µεαβγucγ)(Xµ)aα]
+H.c. . (4.11)
By inspection of the above expression it can be seen that quarks may become leptons (and vice-versa)
in interactions mediated by these bosons. One important consequence of this is that protons can decay.
In order to get an idea of the constraints that the model parameters in (4.11) must obey for there
to be accordance with experimental data it is useful to make an estimation of the proton decay width
associated these terms. Since this is a low energy process, the analysis can be done using effective
43
operators. The equations of motion for a given field φ are
∂µ∂L
∂(∂µφ)− ∂L∂φ
= 0 (4.12)
and when a φ is much heavier than the energy scale of the event being studied, as in the leptoquark
case, the field can be integrated out (its derivatives are very small) so (4.12) is approximately
∂L∂φ
= 0. (4.13)
Before applying equation (4.13) to the leptoquark fields it is still necessary to add the new bosons
mass term,
LMV= −M
2V
2(Xc
µ)αa (Xµ)aα, (4.14)
to (4.11) so that LXY +LMVhas all relevant information of these fields as well as their interactions with
fermions. One can, now, use the equations of motion to get
(Xcµ)αa =
√2g5
M2X
{(dc)αγµεabL
b − ecεbaγµqαb + qβaγµεαβγucγ
}, (4.15)
while the conjugate field is given by the Hermitian conjugate of the previous expression. Substituting in
(4.11) yields
LXeff =g2
5
M2X
εαβγ(uc)αγµqaβ{ecεabγ
µqγb + (dc)γγµεabLb}
+H.c., (4.16)
which are the dimension 6 effective operators.
To get an estimation of the proton decay width associated with (4.16) one can consider the following
decay channel:
p→ π0e+, (4.17)
that is, one up quark becomes a positron and the down quark becomes an anti-up quark. The amplitudes
contributing to this process are
M1 =g2
5
M2V
εαβγ(uc)αuβ ecγµγµdγ , M2 =
g25
M2V
εαβγ(uc)αγµuβ(dc)γγµe, (4.18)
so one can state that [66]
Γpd ∼ α25
m5p
M4X
. (4.19)
Using the most restrictive constraints on the proton lifetime available at this time [67], which are
τ(p→ π0e+) > 1.4× 1034years, (4.20)
and the fact that unification requires 26 ≤ α−15 ≤ 35 [68], we obtain
MV > (4.9− 5.7)× 1015GeV. (4.21)
44
It is being assumed that MV and the unification scale are of the same order of magnitude so (4.21)
constrains the region where gauge couplings can unify. There is enough freedom in the Adjoint model
for these constraints to be respected, therefore, leptoquark mediated proton decay is not a big problem
in this context.
It was pointed out, in the last chapter, that the 5H scalar multiplet includes not only a Higgs-like SU(2)
doublet but also a coloured scalar triplet. Since this triplet is absent in the SM it is important to study
its consequences in the Adjoint model and SU(5) GUTs in general. Its interactions with fermions come
from the Yukawa Lagrangian in (4.2):
− L5H−fermionY =
εαβγδη4
(Γ1u)ij10αβi 10γδj (5H)η +
√2(Γ1
d)ij10αβi 5∗jα(5∗H)β , (4.22)
where the i,j are generation indexes and α, β, γ, δ, η are SU(5) indexes. Some other terms in (4.2) also
involve the 5H and fermions but not quarks so they are not relevant for the present discussion. Taking
into account the fact that 5H may be written as
5H =
T1
H
, (4.23)
where T1 is the coloured scalar triplet and H is the Higgs-like doublet, (4.22) may be simplified and,
keeping only the terms with T1 we get
(Γ1u)ij(
1
2QiQj + uiCejC)T1 + (Γ1
d)(QiLj + uiCdjC)T ∗1 . (4.24)
After adding the triplet’s mass term,
−m2T1T ∗1 T1, (4.25)
to (4.24), we can repeat the procedure used to obtain the effective operators for leptoquark mediated
proton decay to get(Γ1u)ij(Γ
1d)kl
M2T1
[1
2(QiQj)(QkLl) + (ucie
cj)(u
ckdcl )
], (4.26)
the dimension 6 effective operators contributing to the decays by means of T1 bosons exchange.
There is another coloured scalar triplet in the Adjoint model, T2, which belongs to the 45H scalar
representation. Starting form the Yukawa Lagrangian and retaking the steps already described for the
leptoquarks and T1 one arrives at the following effective operators:
4(Γ2u)ij(Γ
2d)kl
M2T2
(uciecj)(u
ckdcl ). (4.27)
The 45H multiplet contains other coloured representations that may lead to proton decay, specifically,
the (3∗, 1, 4/5) triplet and the three triplets ∆−1/3, ∆2/3 and ∆−4/3 (the superscript refers to the electric
charge), which belong to the (3, 3,−1/5) triplet-triplet. By expanding the Lagrangian terms involving
these fields one can see that (3∗, 1, 4/5) and ∆−4/3 make no contribution for proton decay as they only
couple to pairs of up-quarks while (Γ2u) is antisymmetric (therefore it cannot couple two equal particles).
45
For the ∆−1/3 and ∆2/3 fields the effective operators are
(Γ2u)ij(Γ
2d)kl
2M2∆−1/3
[(uidj)(ukel) + (uidj)(dkνl)] (4.28)
and
− (Γ2u)ij(Γ
2d)kl
2M2∆2/3
(didj)(ukνl) (4.29)
respectively.
All sources of proton decay in the Adjoint model are now accounted for. One could approach the
scalar mediated case in the same way as the leptoquark mediated case, by estimating the decay widths
associated with it and, consequently, the constraints imposed on the masses of these scalars, but an-
other strategy will be used. The main reason for this is that, unlike the leptoquark masses, the scalar
masses cannot be raised above a certain threshold without creating or worsening doublet-triplet splitting
problems. Therefore, we will try to suppress scalar mediated proton decay in a different way. Since
these processes depend on the Yukawa matrices of the model, this suppression may be achieved by
imposing constraints on these matrices. However, even if we forbid these processes at tree-level and
take into account the fact that scalar-mediated decays are proportional to the Yukawa terms which are,
in general, smaller than gauge couplings, we cannot eliminate the splitting problems. This endeavour is
still worthwhile because it loosens the constraints on scalar triplet masses.
4.3 Conditions for Proton Decay Suppression and Discrete Sym-
metries
By inspection of the effective operators for scalar mediated proton decay it becomes apparent that this
possess can be eliminated (at tree level) if certain elements of the Yukawa matrices are zero. These
matrices are arbitrary and complex, so it is convenient to find a way of ”forcing” specific elements to
be annulled. This can be done by changing the gauge group to SU(5) × A where A is a discrete
abelian group and assigning A charges to the fields in such a way that some terms are not invariant
for transformations from this group and, therefore, cannot be present. It is important to note that this
discrete symmetry survives below unification scale only if the field responsible for the SU(5) SSB has
no A group charge.
The first step that must be taken in order to achieve this is to identify the conditions that the Yukawa
matrices must verify for proton decay to be forbidden at tree level. Starting form the effective operators
obtained in the previous section one can draw the diagrams that contribute to this process and infer
which Yukawa parameters have to be zero. For T1 mediated proton decay these operators are given in
(4.26) and the corresponding diagrams are shown in Figure 4.1.
The conditions that are to be imposed on Γ1u and Γ1
d are summarized in Table 4.1.
It is important to note that decays with third generation fermions (other than neutrinos) or charmed
quarks in the final state are already forbidden as these particles are heavier than protons and to re-
46
Figure 4.1: Diagrams associated with T1 mediated proton decay. Some diagrams represent severalprocesses as Q and L can correspond to different particles.
Diagram A B C D
Γ1u-exclusive conditions (Γ1
u)11 = 0(Γ1u)11 = 0,
(Γ1u)12 = 0
(Γ1u)11 = 0,
(Γ1u)12 = 0
(Γ1u)11 = 0,
(Γ1u)12 = 0
Γ1d-exclusive conditions
(Γ1d)11 = 0,
(Γ1d)12 = 0,
(Γ1d)21 = 0,
(Γ1d)13 = 0,
(Γ1d)23 = 0,
(Γ1d)22 = 0
(Γ1d)11 = 0,
(Γ1d)12 = 0,
(Γ1d)13 = 0
(Γ1d)11 = 0
(Γ1d)11 = 0,
(Γ1d)12 = 0
Table 4.1: Sufficient conditions for proton decay suppression in the T1 mediated case.
member that, by definition, Γ1u is symmetric and Γ2
u is antisymmetric. Since the diagrams in Fig. 4.1 are
the only ones involving Γ1u and Γ1
d we can already list the distinct, independent, sets of conditions these
matrices must obey in order for our goal to be achieved:
→ (Γ1u)11 = 0, (Γ1
d)11 = 0, (Γ1d)12 = 0, (Γ1
d)13 = 0, (4.30)
→ (Γ1u)11 = 0, (Γ1
u)12 = 0, (4.31)
→ (Γ1d)11 = 0, (Γ1
d)12 = 0, (Γ1d)21 = 0, (Γ1
d)22 = 0, (Γ1d)13 = 0, (Γ1
d)23 = 0. (4.32)
For T2 mediated proton decay we use (4.27) to obtain the diagrams shown in Figure 4.2.:
The conditions required for these processes to vanish are presented in table 4.2.:
Repeating these steps for ∆−1/3 mediated proton decay and using (4.28) the relevant diagrams are
47
Figure 4.2: Diagrams associated with T2 mediated proton decay.
Diagram E F
Γ2u-exclusive conditions (Γ2
u)11 = 0,(Γ2u)12 = 0
(Γ2u)11 = 0,
(Γ2u)12 = 0
Γ2d-exclusive conditions (Γ2
d)11 = 0(Γ2d)11 = 0,
(Γ2d)12 = 0
Table 4.2: Sufficient conditions for proton decay suppression in the T2 mediated case.
Figure 4.3: Diagrams associated with ∆−1/3 mediated proton decay.
and the sufficient conditions for them to vanish are
Finally, regarding ∆2/3 induced proton decay, we can use (4.29) to obtain the remaining proton decay
diagram:
48
Diagram G H I J
Γ2u-exclusive conditions (Γ2
u)11 = 0(Γ2u)11 = 0,
(Γ2u)12 = 0
(Γ2u)11 = 0
(Γ2u)11 = 0,
(Γ2u)12 = 0
Γ2d-exclusive conditions (Γ2
d)11 = 0,(Γ2d)12 = 0
(Γ2d)11 = 0,
(Γ2d)12 = 0
(Γ2d)11 = 0,
(Γ2d)12 = 0,
(Γ2d)21 = 0,
(Γ2d)22 = 0,
(Γ2d)23 = 0
(Γ2d)11 = 0,
(Γ2d)12 = 0,
(Γ2d)13 = 0
Table 4.3: Sufficient conditions for proton decay suppression in the ∆−1/3 mediated case.
Figure 4.4: Diagrams associated with ∆2/3 mediated proton decay.
This process is forbidden if
Diagram K
Γ2u-exclusive conditions (Γ2
u)11 = 0,(Γ2u)12
Γ2d-exclusive conditions
(Γ2d)11 = 0,
(Γ2d)12 = 0,
(Γ2d)13 = 0
Table 4.4: Sufficient conditions for proton decay suppression in the ∆2/3 mediated case.
Combining the results from tables 4.2, 4.3 and 4.4 we can compile the conditions Γ2u and Γ2
d must
verify to suppress proton decay:
→ (Γ2u)11 = 0, (Γ2
d)11 = 0, (Γ2d)12 = 0, (Γ2
d)13 = 0 (4.33)
→ (Γ2u)11 = 0, (Γ2
u)12 = 0 (4.34)
→ (Γ2d)11 = 0, (Γ2
d)12 = 0, (Γ2d)21 = 0, (Γ2
d)22 = 0, (Γ2d)13 = 0, (Γ2
d)23 = 0. (4.35)
If one of (4.30), (4.31) or (4.32) and one of (4.33), (4.34) or (4.35) is true then there is no scalar mediated
proton decay at tree level in the Adjoint model. We now have the right setup to start looking for discrete
abelian symmetries that can lead to these conditions being satisfied.
This was achieved by writing a computer program in the C language that, for a given symmetry, tests
all possible charge assignments that the fields involved in proton decay may have to check weather the
constraints discussed in the previous paragraph are obeyed. The symmetries that were considered are
49
the Z1, Z2, Z3, Z4, Z5, Z2×Z2, Z2×Z3, Z2×Z4 and Z3×Z3. No larger Zn or Zn×Zm cases were used
because they are more difficult to work with from a computational point of view and the solutions they
provide are practically the same as the ones obtained for the studied symmetries (this is corroborated
by tests performed with symmetries up to Z12 and Z5 × Z5). The continuous abelian group was also
included in the analysis for reasons that will be explained in the beginning of the next chapter. Additional
to finding solutions of the proton decay problem, the aforementioned program also uses (3.48), (3.49)
and (3.50) to identify which elements of the fermion mass matrices are zero. Therefore, these solutions
are grouped according to the number of zeros they imply on Mu and Md. The results obtained are
presented in the form of tables that can be found in Appendix C. There is one table for each symmetry
and their elements can be null if there is no charge assignment for which there is a solution and Mu, Md
have the specified amount of zeros or a non-null number that represents the possible different solutions
with the specified amount of mass matrix zeros where the textures and the symmetries displayed by the
Yukawa matrices at GUT scale are taken into account. As an example, two of these tables that are of
relevance for the rest of this work are shown.
Z4
0 1 2 3 4 5 60 4 0 0 0 0 0 01 4 6 12 0 3 0 02 0 0 0 0 0 0 03 0 0 0 0 0 0 04 0 0 0 42 138 54 05 0 0 0 30 105 78 06 0 0 0 0 0 0 6
Table 4.5: The first column indicates the number of Mu zeros and the first row indicates the number ofMd zeros.
Continuous symmetry
0 1 2 3 4 5 60 0 0 0 0 0 0 01 4 6 12 0 3 0 02 0 0 0 0 0 0 03 0 0 0 0 0 0 04 0 0 0 42 111 54 05 0 0 0 0 27 84 426 0 0 0 0 24 96 78
Table 4.6: The first column indicates the number of Mu zeros and the first row indicates the number ofMd zeros.
50
Chapter 5
Adjoint SU(5) with Discrete Symmetry
Results
In order to obtain a realistic theory from the model described in the previous chapter we review aspects
of low-energy fermion masses and mixings that this theory should be able to reproduce. We also see
that the discrete symmetry used to prevent scalar mediated tree level proton decay may be broken by
quantum gravity effects unless some ”discrete anomaly” cancellation conditions are verified. Finally, we
investigate the possibility of choosing a particular ZN symmetry and assigning charges related to this
group in such a way that these constraints are obeyed, while trying to make the resulting theory as
predictive as possible.
5.1 Criteria for a Realistic Adjoint SU(5)×ZN theory
Estimates of the unification scale put it at much higher energies than the ones accessible to us in
controlled experiments. Because of this, it is impossible to directly check certain relations that are valid
when the GUT gauge group is effective or detect the new heavy fields predicted to exist, but we can make
sure that the theory is consistent with low-energy data, determine unification and proton constraints on
the masses of new fields to find out weather some of them may be detected in the near future and see
if there are new (accurate) relations between low-energy observables not present in the SM.
We start by verifying the theory’s compatibility with low-energy data, for if this fail it loses most of its
interest. In Chapter 2. we introduced the CKM matrix, which contains all the information about quark
mixings at SM scale, so we should verify if there are enough free (quark related) Yukawa parameters
to get this matrix, within 1σ of its experimental values, at such energies. When we talked about quark
mixings it was pointed out that no analogous processes took place in the lepton sector. This was
justified by the absence of neutrino masses in the SM, but, since then, we have managed to generate
light masses for left-handed neutrinos through seesaw mechanisms.
51
5.1.1 Low-Energy Neutrino Phenomenology
Experimental observation of neutrino oscillations strongly suggests that these particles have masses (as
it depends on the square of different neutrino’s masses) and implies a distinction between the interaction
and mass basis, specifically,
ν′α =
3∑j=1
Uαjνj , (5.1)
where U is an unitary matrix, ν′ are interaction eigenstates and the ν are mass eigenstates. After SSB
of the GUT gauge group we may write the lepton masses in the interaction basis as
Lleptonmasses = −Mle′Le′R −
1
2mµν
′Lν′cL +H.c.. (5.2)
The charged lepton mass matrix can be diagonalized by means of a bi-unitary transformation, that is
U†lLMlURl = Dl ≡ diag(me,mµ,mτ ). (5.3)
As for the neutrino Majorana mass matrix, it is symmetric and is multiplied form both sides by the same
vector (in family space), νL, so we diagonalize it using a single unitary matrix,
UTν mνUν = Dν ≡ diag(m1,m2,m3), (5.4)
where the mi represent left-handed neutrino masses. Interaction and mass eigenstates are, then, re-
lated through
l′R = URllR, l′L = ULllL, ν
′L = UννL. (5.5)
Applying these transformations to the charged current in (2.42) we obtain
Lchargedcurrent =gW√
2(u′Lγ
µd′L+ν′Lγµe′L)W+
µ H.c. =gW√
2(uLVCKMγ
µdL+νLU†PMNSeL)W+
µ +H.c., (5.6)
where
UPMNS = U†LlUν =
Ue1 Ue2 Ue3
Uµ1 Uµ2 Uµ3
Uτ1 Uτ2 Uτ3
. (5.7)
For the same number of generations PMNS has more phases than VCKM , as long as the neutrinos
are Majorana particles. This happens because we cannot rephase Uν without altering the neutrino
masses:
Uν → Uνdiag(eiξ1 , eiξ2 , eiξ3)⇒ mj → mje2iξj . (5.8)
We conclude that ξj can only be arbitrary if mj = 0. The PMNS matrix may be written as
UPMNS = diag(eiζe , eiζµ , eiζτ )Uinternaldiag(eiβ1 , eiβ2 , eiβ3), (5.9)
52
where Uinternal has the same number of parameters as VCKM , with ng(ng − 1)/2 Euler angles and
(ng − 1)(ng − 2)/2 Dirac phases. Considering ng = 3, a frequently used way of parametrizing Uinternal
is
Uinternal =
c12c13 s12c13 s13e
−iδ
−s12c23 − c12s23s13eiδ c12c23 − s12s23s13e
iδ s23c13
s12s23 − c12c23s13eiδ −c12s23 − s12c23s13e
iδ c23c13
, (5.10)
with ci ≡ cos θi and si ≡ sin θi for i = 12, 13, 23. We note that all ζα phases are devoid of physical
meaning as they may be eliminated by rephasing the charged lepton fields. This procedure is not
applicable to the Majorana phases βj due to already mentioned reasons. However, only two out of these
three phases (with ng = 3 and no zero neutrino mass) are relevant because we can remove the global
phase (β1 + β2 + β3)/3 and add it to the ζ phases. In fact, the potentially observable quantities involving
βs are 2(β2 − β1) and (β3 − β1).
Then, for any given number of generations, UPMNS has the same number of phases as VCKM plus a
number of Majorana phases equal to the number of Majorana neutrinos minus one, unless one of these
neutrinos has a null mass (experimental data collected so far is consistent with one massless neutrino),
in which case the number of Majorana phases is reduced by one. The UPMNS matrix is very important
in this context as it provides us a way of testing low-energy lepton mixings and neutrino phenomenology
given by a GUT.
5.1.2 Discrete Gauge Symmetry Anomalies
A QFT with a gauge group of SU(5) and a fermion content consisting in ng 5∗, ng 10 and nρ adjoint (24)
representations is anomaly-free. The only possible anomaly in this scenario would be the [SU(5)]3 one
but 5∗ and 10 have opposite contributions while 24 representations do not contribute at all [69]. What
about the SU(5)× ZN case?
At first sight, one would think that no new anomalies arise, as discrete symmetries are not, in general,
constrained by anomaly-freedom conditions. However, it has been argued that quantum gravity effects
violate any non-gauge symmetries. Then, if we want our discrete symmetry to be safe from these effects,
we must make sure they verify certain ”discrete anomaly cancellation” conditions [70]. In order to define
such conditions we note that in continuum theories the ZN discrete symmetries result from breaking an
U(1) gauge symmetry with a scalar field φ of charge Nq that acquires a vev. Therefore, the ”discrete
anomaly” constraints correspond to the requirement that the original U(1) group is anomaly-free.
If we have a discrete ZN gauge symmetry under which the fermion fields transform as Ψi →
exp(i2πqi/N)Ψ1, the U(1) charges before SSB were necessarily of the form
Qi = qi +miN, (5.11)
where qi and mi are integers (qi are the ZN charges). For an SU(5) × U(1) gauge group the possible
53
anomalies are U(1) [SU(5)]2, [U(1)]
3 and U(1)-gravity-gravity, which are proportional to
A1 =∑i
Ti(Qi), (5.12)
where Ti are the quadratic Casimirs,
A2 =∑i
(Qi)3, (5.13)
and
A3 =∑i
Qi, (5.14)
respectively (the sum is over the theory’s fermion representations). We may use (5.11) to redefine the
U(1) charges so these As are annulled, but this is not always possible. In particular, note that, when
using the ZN charges, the As must be proportional to N for this procedure to work (necessary but not
sufficient condition).
We can also consider the existence of heavy fermions that become massive upon SSB and con-
tribute to the anomaly-freedom conditions. There are two types of masses fermions may acquire at this
stage [71]:
-Pairs of different Weyl fermions with charges that verify
QjW1 +QjW2 = pjN, (5.15)
where pj is an integer, combine to get a mass. They contribute an amount of
Tjpj (5.16)
to A1, an amount of
(QjW1)3 + (QjW2)3 = pjN(3Qj2W1 − 3pjNQjW1 + p2
jN2) (5.17)
to A2 and an amount of
pjN (5.18)
to A3.
-One fermion that is an SU(5) singlet gets a Majorana mass and it must verify
Qjχ =1
2p′jN (5.19)
when N is even (there is no contribution to the anomalies when N is odd), pj are integers. A1 is left
unchanged because the Casimir of an SU(5) singlet is zero, A2 is modified by the addition of
1
8p′3j N
3 (5.20)
54
and A3 by the addition of1
2p′jN. (5.21)
Admitting the possibility of having several of these new fields and considering the charge redefinitions
we are free to make, the anomaly-freedom constraints become
∑i
Ti(qi) = rN, (5.22)
∑i
(qi)3 = mN + ηn
N3
8, (5.23)
and ∑i
qi = pN + ηqN
2, (5.24)
where η is 1, 0 for N even, odd, respectively, and the integers r, m, n, p, q are given by
r = −∑i
Timi −∑
heavies
Tjpj , (5.25)
m = −(3∑i
q2imi + 3N
∑i
qim2i +N2
∑i
m3i )
−∑j
pj(3Qj2W1 − 3pjNQ
jW1 + p2
jN2),
n =∑j
p′3j ,
(5.26)
p = −∑i
mi −∑j
pj , q = −∑j
p′j . (5.27)
If the conditions in (5.22), (5.23) and (5.24) are satisfied we still have no guarantee of anomaly-
freedom, since (5.25), (5.26) and (5.27) are diophantine equations that do not always admit solutions.
As we will see later, the new heavy fields introduced in this section may help us achieve a theory without
anomalies, but it can be argued that carefully choosing the characteristics and number of fields to be
added in order to achieve this goal is not natural, so, in this context, the obtained conditions are more
useful in excluding theories that cannot be made anomaly-free (unless some other mechanism comes
into play).
5.1.3 Unification Constraints
An analysis of unification-related constraints on the model’s fields masses can be carried out before
a particular ZN is chosen because the particle content will be the same and tree-level scalar-mediated
proton decay suppression is required in all studied cases (this is not exactly true as relationships between
the masses of the adjoint fermionic fields depend on (4.3), but it has been checked that this does not
significantly affect our discussion, we consider an M with all elements equal to zero except for the ones
in the diagonal for simplicity). From this point onward we consider nρ = 3 as a smaller number could
55
lead to difficulties in adequately reproducing low-energy neutrino phenomenology [68].
Before performing computations to obtain the field masses compatible with unification it is impor-
tant to take into account the constraints associated with proton decay. As previously deduced, the
leptoquarks must have masses of the same order as the GUT scale in order to prevent these decay
processes, so, recalling (3.14) and (3.15) we conclude that they do not contribute to gauge coupling
runnings. Similar considerations impose mT1,2≥ 3× 1011GeV [66] and m∆ ≥ 1.2× 1013GeV [72]. How-
ever, since the ZN discrete symmetry forbids scalar-mediated decays at tree-level, these constraints are
alleviated.
Bearing this in mind, the relevant contributions to the B-test in our Adjoint SU(5) × ZN model are
shown in Table 5.1. It can be seen that the only fields contributing towards unification (increasing the B
value in relation to the SM one) are the Higgs doublets, ρ3, ∆ and Σ3. From considerations made in the
previous chapter, we should have a large mass for ρ8 and we also assume mρ8� mρ3
, which favours
unification. Regarding the Higgs doublets, they are expected to have masses around the electroweak
scale, so rH1,2 ' 1. Due to their negative effect on unification (except for ∆) and proton decay constraints
(involving ∆) we take the masses of the 45H scalar fields to be of the unification scale order, so they
have no influence on gauge coupling runnings. Although they have less constrained masses than the
leptoquark or ∆, the scalar triplets T1,2 decrease the B-test value, therefore, we also assign them masses
around the unification scale.
S(8,2) S(6∗,1) S(3∗,2) S(3∗,1) ∆ T1,2 H1,2
B23 − 23rS(8,2)
− 56rS(6∗,1)
16rS(3∗,2)
− 16rS(3∗,1)
23r∆ − 1
6rT1,2
16rH1,2
B12 − 815rS(8,2)
215rS(6∗,1)
1715rS(3∗,2)
1615rS(3∗,1)
− 95r∆
115rT2
− 115rH2
Σ8 Σ3 ρ8 ρ(3,2) ρ3
B23 − 12rΣ8
13rΣ3
−2rρ8
13rρ(3,2)
43rρ3
B12 0 − 13rΣ3
0 23rρ(3,2)
− 43rρ3
Table 5.1: Relevant contributions to the B-test in the Adjoint SU(5)× ZN model.
The masses we study in order to see weather or not unification may occur are, the, the ones as-
sociated with the adjoint fermion fields and the Σ8, Σ3 scalars. Since these last two masses are not
constrained by proton decay we allow them to vary over several orders of magnitude, but it should be
noted that by doing this we are ignoring naturality arguments that were already discussed in the pre-
56
vious chapter. To simplify our computations, we assume that two mass states of the adjoint fermionic
representations are degenerate, while keeping in mind the fact that, for the seesaw mechanisms to work
properly, all these masses should have high values (in comparison to the electroweak scale). Finally,
to further simplify this analysis we consider that the ρ(3,2) fields have large enough masses for their
contributions to negligible.
The most immediate conclusion we can reach by investigating gauge coupling runnings in this context
is that unification is easily achieved. In fact, there is a large range of values for the masses mentioned
in the last paragraph that lead to this. Furthermore, unification may occur for different energy scales,
specifically, for 26 ≤ α−1U ≤ 35. The main reason behind all this freedom for the masses being studied
is the inclusion of 3 adjoint fermion representations. For example, the Σ8 and Σ3 masses can go from
the electroweak to the GUT scales or, alternatively, if we fix these values using naturality arguments, the
masses of the scalars in 45H (with exception of ∆) may also vary between those thresholds. All that was
stated in this paragraph is verified even when the scalars that mediate proton decay have the smallest
mass allowed by constraints related to this phenomenon. In conclusion, we clearly have a GUT in this
setup, consistent with a wide variety of possible spectra for the new fields that are not present in the SM,
including cases where some of these masses are close to the electroweak scale, which means that they
may be detected in the near future. This, however, comes at the cost of a larger number of adjustable
parameters in the model and, as a consequence, we are unable to find correlations between the masses
being studied.
5.2 Study of two particular solutions
In this section we analyse two particular realizations of the Adjoint SU(5) × ZN model by checking
weather or not they are compatible with experimental data and by investigating their predictivity. First,
we study a case with N = 8 that suppresses tree-level scalar-mediated proton decay and gives up-
,down-quark matrices with 5 and 3 texture zeros, respectively. The reason for this choice has to do with
the fact that, as we saw in the last chapter(or in Appendix C), this textures cannot be obtained using a
continuous symmetry. If this was not the case it could be argued that we are actually working with a
continuous symmetry and not a discrete one. There are other possible textures verifying that, but this
choice appears to have an adequate balance between enough free parameters to fit low-energy data and
not so many of them that no new correlations among physical quantities can be obtained. In the second
case we have N = 7 and this symmetry leads to up-, down-quark matrices in the Nearest Neighbour
Interaction form, which also guarantees the absence of tree-level scalar-mediated proton decay.
57
5.2.1 Z=8
There are many Z8 charge assignments that provide the texture we are interested in. The one that was
chosen corresponds to
Q(101) = 2, Q(102) = 0, Q(103) = 4, Q(5∗1) = 6, Q(5∗2) = 0, Q(5∗3) = 6, Q(5H) = 0,
Q(45H) = 2,(5.28)
the reason to pick this particular combination will become clear later. From (5.28) and (4.2) it follows that
Γ1u =
0 0 0
0 ∗ 0
0 0 ∗
, (5.29)
Γ1d =
0 ∗ 0
0 0 0
∗ 0 ∗
, (5.30)
Γ2u =
0 0 ∗
0 0 0
∗ 0 0
, (5.31)
Γ2d =
∗ 0 ∗
0 ∗ 0
0 0 0
, (5.32)
where ∗ represents arbitrary complex numbers.
Using (3.53) and (3.54) one may, now, write
Mu =
0 0 ∗
0 0 0
∗ 0 ∗
(5.33)
and
Md =
∗ ∗ ∗
0 ∗ 0
∗ 0 ∗
. (5.34)
Recalling the definitions of Hu and Hd given in Chapter 2 we get
Hu = K† ∗
A 0 B
0 C 0
B 0 D
∗K (5.35)
58
and
Hd = K ′† ∗
A′ B′ C ′
B′ D′ 0
C ′ 0 E′
∗K ′ (5.36)
where A, B, C, D, A′, B′, C ′, D′ and E′ are real, positive numbers and K, K ′ are diagonal matrices of
the form
K,K ′ =
eα1,α2 0 0
0 eβ1,β2 0
0 0 1
, (5.37)
α and β are phases.
In order to determine the CKM matrix one must diagonalize Hu and Hd. Taking (2.46) into account
we may state that
V = OTuK′′Od, (5.38)
where Ou is an orthogonal matrix that diagonalizes Hu, Od is an orthogonal matrix that diagonalizes Hd
and K ′′ = K† ∗K ′. These orthogonal matrices are going to be obtained by resorting to computational
methods, but before doing so it is convenient to find the regions where the parameters represented by
capital letters have to be so that the H ’s contain the correct information.
The H matrices must, upon being diagonalized through bi-unitary transformations, return diagonal
matrices with the squares of the quark masses as their elements. We can take advantage of the fact
certain matrix properties remain invariant for such transformations to write equations relating the afore-
mentioned parameters. The invariances of the trace, determinant and χ yield
tu = A+ C +D, (5.39)
du = A(CD −B2), (5.40)
χu = AC +AD + CD −D2, (5.41)
for Hu with tu = m2u +m2
c +m2t , du = m2
u ×m2c ×m2
t , χu = m2u ×m2
c +m2u ×m2
t +m2c ×m2
t and
td = A′ + C ′ + E′, (5.42)
dd = A′(D′E′ − C ′2)−B′2E′, (5.43)
χd = A′D′ +A′E′ +D′E′ −B′2 − C ′2, (5.44)
for Hd with td = m2d +m2
s +m2b , dd = m2
d ×m2s ×m2
b , χd = m2d ×m2
s +m2d ×m2
b +m2s ×m2
b , respectively.
These equations can be used to get A = tu − r3 and B, C in terms of D:
C = r3 −D, (5.45)
59
B2 = (tu − r3)r3 +D(r3 −D)− χu, (5.46)
where r3 is the third root of x3 − 2tux2 + (t2u + χu)x + du − tuχu = 0. Since all these parameters must
be greater than zero D has to verify
−D2 + r3D + tur3 − r23 − χu > 0. (5.47)
As for the Hd parameters, they may be written in terms of D′ and E′:
A′ = td −D′ − E′, (5.48)
B′2 =A′χd − dd −A′2(D′ + E′)
E′ −A′, (5.49)
C ′2 = A′D′ +A′E′ +D′E′ − χd −B′2. (5.50)
The requirement that all these quantities be greater than zero leads, when E′ > A′, to
y3 − 2tdy2 + (t2d + χd)y + dd − tdχd < 0 (5.51)
and
E′3 − tdE′2 + χdE′ − dd < 0 (5.52)
with y = D′ + E′, while E′ < A′ leads to
y3 − 2tdy2 + (t2d + χd)y + dd − tdχd > 0 (5.53)
and
E′3 − tdE′2 + χdE′ − dd < 0. (5.54)
Using these conditions and an appropriate computer program (in C language) an attempt was made
to discover values for D, D′, E′ and the phases in K ′′ for which the VCKM matrix obtained would be
consistent with experimental data within one standard deviation. This attempt was successful and the
resulting data was collected for further analysis. To check for correlations between VCKM quantities
several plots were made. These plots are shown in Figures 5.1 to 5.5:
It can be seen that, for most plots, the results obtained in this work coincide with the available
parameter space, that is, there are no correlations. The exception is the J vs. sin 2β plot, where some
correlation seems to be present.
Before moving on to the neutrino and charged lepton phenomenology, it is important to note that,
because of the charges assigned to 5H and 45H , all terms in the scalar potential (see Appendix B)
involving simultaneously 5H , 24H and 45H must be removed as they are not Z8-invariant. This leads
to an accidental global continuous symmetry which, upon electroweak SSB, would lead to a massless
Nambu-Goldstone boson at tree-level [73]. We can avoid this problem by adding an SU(5) singlet scalar
field S1 with a Z8 charge of 6, which allows terms of the form 5∗H45HΣS1 + H.c. to be present in the
60
Figure 5.1: Correlation plots involving VCKM moduli, specifically, Vub vs. Vcb. The red dots representthis work’s results, the blue ones represent all the possibilities.
Figure 5.2: Correlation plots involving VCKM moduli, specifically, Vus vs. Vcb. The red dots representthis work’s results, the blue ones represent all the possibilities.
scalar potential.
From (3.55) and the Yukawa matrices we have already determined, we get
Me =
∗ 0 ∗
∗ ∗ 0
∗ 0 ∗
. (5.55)
61
Figure 5.3: Correlation plots involving VCKM moduli, specifically, Vus vs. Vub. The red dots representthis work’s results, the blue ones represent all the possibilities.
Figure 5.4: Correlation plots involving VCKM angles, specifically, γ vs. sin 2β . The red dots representthis work’s results, the blue ones represent all the possibilities.
As for the effective light neutrino mass matrix generated through type I and type III seesaw, it depends
on the Z8 charges assigned to the adjoint fermion representations. In this case we chose Q(ρ1) =
62
Figure 5.5: Correlation plots involving VCKM angles, specifically, J vs. sin 2β. The red dots representthis work’s results, the blue ones represent all the possibilities.
5, Q(ρ2) = 6, Q(ρ3) = 2, which leads to
Γ1ν =
0 0 0
0 ∗ 0
0 0 0
, (5.56)
Γ2ν =
0 0 ∗
0 0 0
0 0 ∗
(5.57)
and
M = λ =
0 0 0
0 0 ∗
0 ∗ 0
. (5.58)
Since the adjoint fermion fields can couple to the scalar singlet S1, that may acquire a vev, the total
Majorana mass matrix receives contributions from
MS1=
∗ 0 0
0 0 0
0 0 0
, (5.59)
but this is not enough to get a realistic effective light neutrino mass matrix, so we add another SU(5)
63
singlet scalar S2 with an Z8 charge of 5, yielding
MS2=
0 ∗ 0
∗ 0 0
0 0 0
(5.60)
and, consequently,
MTotal =
∗ ∗ 0
∗ 0 ∗
0 ∗ 0
. (5.61)
Now, using (4.7) and (4.6) we obtain
(mν)ij =
∗ ∗ ∗
∗ 0 ∗
∗ ∗ ∗
. (5.62)
Taking into account the studies of neutrino textures in ref. [74], we conclude that this case is compatible
with experimental results, although it provides little to no predictivity. Making different charge assign-
ments it is possible to have the same up-, down-quark textures and tree-level scalar-mediated proton
decay suppression with a different effective neutrino mass matrix. The viable possibilities include other
textures with only one zero or textures with three independent zeros. It would be interesting to inves-
tigate these last cases as they may offer some predictivity (provided they are consistent with current
low-energy data).
Finally, we check the ”discrete gauge anomaly” conditions. Using Z8 charges, the factors in (5.12),
(5.13) and (5.14) are
A1 = 160, (5.63)
A2 = 11256 (5.64)
and
A3 = 432. (5.65)
It is easy to see that all these factors are proportional to 8, but it is not possible to annul them using
only the ambiguity in the definition of U(1) charges. Instead, we take advantage of this freedom to make
A1 = 0. Choosing the U(1) charges as
Q′(101) = 18, Q′(102) = 40, Q′(103) = 36, Q′(5∗1) = −34, Q′(5∗2) = −40, Q′(5∗3) = −18,
Q′(ρ1) = 13, Q′(ρ2) = −2, Q′(ρ3) = −30,
(5.66)
we get
A′1 = 0, (5.67)
64
A′2 = 23736 (5.68)
and
A′3 = 24. (5.69)
Recalling (5.20) and (5.21) we note that, if (5.68) was proportional to N3/8 = 64 and (5.69) was propor-
tional N/2 = 4, we could try to get rid of the anomalies by adding new heavy Majorana fields. However,
(5.68) does not verify this condition, we have, instead, (A′2 − 56) ∝ 64. Then, we may attempt to use
(5.17) by adding pairs of heavy Weyl fermions to the theory in order to get a −56 contribution in A′2.
Considering the existence of four new pairs of such fermions and that they are singlets with repect to the
SU(5) gauge group, we obtain the desired result if three of those fields have pj = 1 and QjW1 = 2 while
the other has pi = −1 and QiW1 = 1. The effect of these additions in A′3 is given by N ×∑j
pj = 16. We
are left with A′223680 and A′3 = 40. There is no integer p′j for which 64× p′3j = 23680 and 4× p′j = 40 but,
introducing two new heavy Majorana fields with −3 and −7 U(1) charges we see that ”discrete gauge
anomaly” freedom is achieved. The procedure carried out to get this result is, as previously stated, not
very natural. Despite this, what is important to take retain is that we can prove that our discrete symmetry
may be protected from quantum gravity effects when possible higher energy physics is considered.
5.2.2 Z=7
A Z4 discrete symmetry is enough to obtain quark mass matrices in NNI form and interesting neutrino
mass matrix textures [68]. We useN = 7 because this is the smallest value ofN for which the necessary
”discrete gauge anomaly” conditions (5.22), (5.23) and (5.24) can be verified. With this in mind, the Z7
charges are (among other possibilities):
Q(101) = 1, Q(102) = 3, Q(103) = 2, Q(5∗1) = 6, Q(5∗2) = 1, Q(5∗3) = 0, Q(5H) = 3,
Q(45H) = 2.(5.70)
Following the same lines as in the previous case,
Γ1u =
0 ∗ 0
∗ 0 0
0 0 ∗
, (5.71)
Γ1d =
0 0 0
0 0 ∗
0 ∗ 0
, (5.72)
Γ2u =
0 0 0
0 0 ∗
0 ∗ 0
, (5.73)
65
Γ2d =
0 ∗ 0
∗ 0 0
0 0 ∗
(5.74)
and the NNI form is given by
Mu =
0 0 ∗
0 0 0
∗ 0 ∗
(5.75)
and
Md =
∗ ∗ ∗
0 ∗ 0
∗ 0 ∗
. (5.76)
Instead of investigating the quark mixings, as we did in the last section, we will focus on the low-
energy neutrino phenomenology. With (5.70) and choosing Q(ρ1) = 4, Q(ρ2) = 5, Q(ρ3) = 6, we have
Γ1ν =
0 ∗ 0
0 0 0
∗ 0 0
, (5.77)
Γ2ν =
0 0 ∗
∗ 0 0
0 ∗ 0
(5.78)
and
M = λ =
0 0 0
0 0 0
0 0 0
. (5.79)
Clearly, we need to add SU(5) singlet scalars to change this situation. Introducing an S1 with a Z7
charge of 4 we get
MS1 =
0 0 ∗
0 ∗ 0
∗ 0 0
, (5.80)
and, consequently,
MTotal =
0 0 ∗
0 ∗ 0
∗ 0 0
. (5.81)
Using (4.7) and (4.6) we obtain
(mν)ij =
∗ ∗ ∗
∗ 0 0
∗ 0 ∗
. (5.82)
66
It should be noted that, in this context, we do not solve the scalar potential problem by adding an-
other scalar singlet, instead we introduce a soft-breaking term:5∗H45HΣ + H.c.. Returning to the neu-
trino physics, we have a texture with two independent zeros. Plugging this result and the form of the
charged lepton mass matrix into a program developed with the minuit2 minimization package for C and
C++ we tested this model’s predictions. The Chi-square test performed returns a minimum value of
7.65627× 10−6, so we conclude that the free Yukawa parameters in the theory are sufficient to satisfac-
torily reproduce experimental data. An analysis of the other results provided by the computer program
will be featured in an addendum that will later be added to this thesis.
67
Chapter 6
Conclusions
In this work, GUTs were introduced as an attempt to solve some of the SM’s problems. Even though the
minimal (SU(5)) GUT failed to solve most of these problems, we saw that it provides a good framework
in which to search for better models. Some extensions, namely, addition of a 45H scalar representation
and seesaw mechanisms allow many problems to be solved and make it possible to actually achieve
gauge coupling unification.
The Adjoint SU(5) model is obtained by combining these extensions. Comparing it with the minimal
model we note that the wrong mass relation predictions are solved, light neutrino masses are generated
and there is plenty of parametric freedom to reach unification. However, these improvements are ac-
companied by an extension of the particle content and consequent increase in the number of adjustable
parameters in the theory. This means that the model becomes less predictive.
One way to improve this situation involves the use of discrete symmetries. They are frequently used
in model building to achieve benefits such as eliminating baryon- or lepton-number violating terms, im-
posing restrictions on the arbitrary Yukawa parameters or suppressing flavour changing neutral currents.
In this work we take advantage of a discrete symmetry to forbid tree-level scalar-mediated proton decay.
This is useful since proton decay events are yet to be detected, therefore, constraints related to these
processes keep getting more stringent.
By imposing a particular discrete symmetry we try to build realistic models and investigate their
predictivity. We see that compatibility with experimental data is easily achieved. As for predictive power,
the chosen neutrino textures were not very promising and other possible choices may give more insights
into relations between physical quantities. Aside from this, several problems still plague these models,
like splitting problems (some of the model’s parameters have to be fine-tuned in order to obtain the
results we need) or the lack of family structure.
The elevated number of fields contributing to gauge coupling runnings means that we can achieve
unification in many different ways but we cannot infer correlations between their masses. On the other
hand, some of these masses may be low enough for the fields to be detected in the near future, which
constitutes an attractive feature. A study of the signature such events would have is beyond the scope
of this work.
69
We have derived sufficient conditions for ”discrete gauge anomaly” freedom to save the discrete
symmetry form quantum gravity effects. It is not always possible obey them but we have seen that the
introduction of certain heavy (above GUT scale) fields can provide help in this context.
Future work in this area may involve the investigation of other discrete symmetries, experimenting
with a different particle content or studying different textures arising from the already used symmetries.
Furthermore, it was assumed in Chapters 4 and 5 that neutrinos are Majorana particles, so it would be
interesting to compare the predictions related to neutrinoless double beta decay [75] with experimental
data to check weather or not this assumption is valid.
70
Bibliography
[1] T. Kinoshita, History of Original Ideas and Basic Discoveries in Particle Physics. Springer, US,
1996.
[2] C. N. Yang and R. L. Mills, “Conservation of isotopic spin and isotopic gauge invariance,” Phys.
Rev., vol. 96, no. 191, 1954. arXiv:1207.7235[hep-ex].
[3] P. W. Higgs, “Broken symmetries and the masses of gauge bosons,” Phys. Rev. Lett., vol. 13,
no. 508-509, 1964.
[4] F. Englert and R. Brout, “Broken symmetry and the mass of gauge vector mesons,” Phys. Rev. Lett.,
vol. 13, no. 321, 1964.
[5] G. Guralnik, C. R. Hagen, and T. W. B. Kibble, “Global conservation laws and massless particles,”
Phys. Rev. Lett., vol. 13, no. 585, 1964.
[6] S. Weinberg, “A model of leptons,” Phys. Rev. Lett., vol. 19, no. 1264, 1967.
[7] A. Salam Conf. Proc., vol. C680519, no. 367, 1968.
[8] S. L. Glashow, “Partial-symmetries of weak interactions,” Nucl. Phys., vol. 22, no. 579, 1961.
arXiv:1207.7235[hep-ex].
[9] G. t’Hooft, “Renormalizable lagrangians for massive yang-mills fields,” Nucl. Phys. B, vol. 35,
no. 167, 1971.
[10] C. Berger, “Observation of a narrow resonance formed in e+e− annihilation at 9.46 gev,” Phys. Lett.
B, vol. 76, no. 243, 1978. arXiv:1207.7235[hep-ex].
[11] G. Arnison, “Experimental observation of isolated large transverse energy electrons with associated
missing energy at√s = 540gev,” Phys. Lett. B, vol. 122, no. 103, 1983. arXiv:1207.7235[hep-ex].
[12] M. Banner, “Observation of single isolated electrons of high transverse momentum in events
with missing transverse energy at the cern pp collider,” Phys. Lett. B, vol. 122, no. 476, 1983.
arXiv:1207.7235[hep-ex].
[13] G. Arnison, “Experimental observation of lepton pairs of invariant mass around 95 gev/c2 at the
cern sps collider,” Phys. Lett. B, vol. 126, no. 398, 1983. arXiv:1207.7235[hep-ex].
71
[14] G. Aad et al., “Observation of a new particle in the search for the Standard Model Higgs boson with
the ATLAS detector at the LHC,” Phys. Lett. B, vol. 716, pp. 1–29, 2012.
[15] S. Chatrchyan et al., “Observation of a new boson at a mass of 125 GeV with the CMS experiment
at the LHC,” Phys.Lett., vol. B716, pp. 30–61, 2012.
[16] H. Georgi and S. Glashow, “Unity of all elementary particle forces,” Phys. Rev. Lett., vol. 32, no. 438-
441, 1974. arXiv:1207.7235[hep-ex].
[17] P. F. Perez, “Renormalizable adjoint su(5),” Phys. Lett. B, vol. 654, no. 189-193, 2007. hep-
ph/0702287.
[18] E. S. Abers and B. W. Lee, “Gauge theories,” Phys. Rept. 9, vol. 9, no. 1, 1973.
[19] H. Georgi, Lie Algebras in Particle Physics. Westview Press, USA, 1999.
[20] D. Bailin and A. Love, Introduction to Gauge Field Theory. Institute of Physics Publishing, Bristol
and Philadelphia, 1993.
[21] J. Goldstone, A. Salam, and S. Weinberg, “Broken Symmetries,” Phys.Rev., vol. 127, pp. 965–970,
1962.
[22] M. E. Peskin and D. V. Schroeder, An Introduction to Quantum Field Theory. Addison-Wesley
Publishing Company, 1995.
[23] N. Cabibbo, “Unitary symmetry and leptonic decays,” Phys. Rev. Lett., vol. 10, no. 531, 1963.
[24] M. Kobayashi and T. Maskawa, “Cp-violation in the renormalizable theory of weak interaction,” Prog.
Theor. Phys., vol. 49, no. 652, 1989.
[25] G. C. Branco, L. Lavoura, and J. P. Silva, CP violation. Clarendon Press-Oxford, 1999.
[26] A. D. Sakharov, “Violation of cp invariance, c asymmetry, and baryon asymmetry of the universe,”
JETP Lett., vol. 5, no. 24, 1967.
[27] M. Gavela, P. Hernandez, J. Orloff, O. Pene, and C. Quimbay, “Standard model CP violation and
baryon asymmetry. Part 2: Finite temperature,” Nucl.Phys., vol. B430, pp. 382–426, 1994.
[28] R. A. Bertlmann, Anomalies in Quantum Field Theory. Clarendon Press-Oxford, 1996.
[29] S. L. Adler, “Axial-vector vertex in spinor electrodynamics,” Phys. Rev., vol. 177, no. 2426, 1969.
[30] J. S. Bell and R. Jackiw, “A pcac puzzle:π0 → γγ in the σ-model,” Nuovo Cim. A, vol. 60, no. 47,
1969.
[31] T. Cheng and L. Li, Gauge theory of elementary particle physics. Clarendon Press, Oxford, 1982.
[32] D. J. Gross and R. Jackiw, “Effects of anomalies on quasi-renormalizable theories,” Phys. Rev. D,
vol. 6, no. 477, 1972.
72
[33] G. t’Hooft and M. J. G. Veltman, “Regularization and renormalization of gauge fields,” Nucl. Phys.
B, vol. 44, no. 189, 1972.
[34] W. Pauli and F. Villars, “On the invariant regularization in relativistic quantum theory,” Rev. Mod.
Phys., vol. 21, no. 434, 1989. arXiv:1207.7235[hep-ex].
[35] A. J. Bruce, Anomalies in Quantum Field Theory. PhD thesis, University of Sussex, 2005.
[36] R. Delbourgo and A. Salam, “The gravitational correction to pcac,” Phys. Lett. B, vol. 40, no. 381,
1972.
[37] H. Georgi and S. L. Glashow, “Gauge theories without anomalies,” Phys. Rev. D, vol. 6, no. 429,
1972.
[38] A. Strumia and F. Vissani, “Neutrino mass mixings and...,” arXiv:hep-ph/0606054, 2010. hep-ph.
[39] E. Golowich and P. B. Pal, “Charge quantization fom anomalies,” Phys. Rev. D, vol. 41, no. 11, 1990.
[40] K. S. Babu and R. N. Mohapatra, “Quantization of electric charge from anomaly constraints and a
majorana neutrino,” Phys. Rev. D, vol. 41, no. 1, 1990.
[41] A. Perez-Lorenzana and W. A. Ponce, “Guts and string guts,” Europhys. Lett., vol. 49, no. 296-301,
2000.
[42] K. Olive et al., “Review of Particle Physics,” Chin.Phys., vol. C38, p. 090001, 2014.
[43] A. Giveon, L. J. Hall, and U. Sarid, “SU(5) unification revisited,” Phys.Lett., vol. B271, pp. 138–144,
1991.
[44] D. Emmanuel-Costa, E. T. Franco, and R. G. Felipe, “Su(5)xsu(5) unification revisited,” JHEP,
vol. 1108, no. 017, 2011. arXiv:1104.2046 [hep-ph].
[45] J. R. Ellis and M. K. Gaillard, “Fermion masses and higgs representations in su(5),” Phys. Lett. B.,
vol. 88, no. 315, 1979. arXiv:1207.7235[hep-ex].
[46] H. Georgi and C. Jarlskog, “A new lepton - quarl mass relation in a unified theory,” Phys. Lett. B,
vol. 86, no. 297-300, 1979.
[47] P. Eckert, J. Gerard, H. Ruegg, and T. Schucker, “Minimization of the SU(5) Invariant Scalar Poten-
tial for the Fortyfive-dimensional Representation,” Phys.Lett., vol. B125, p. 385, 1983.
[48] P. Kalyniak and J. N. Ng, “Symmetry Breaking Patterns in SU(5) With Nonminimal Higgs Fields,”
Phys.Rev., vol. D26, p. 890, 1982.
[49] I. Dorsner, S. Fajfer, J. F. Kamenik, and N. Kosnik, “Light Colored Scalar as Messenger of Up-Quark
Flavor Dynamics in Grand Unified Theories,” Phys.Rev., vol. D82, p. 094015, 2010.
[50] S. Weinberg, “Baryon - and lepton - nonconserving processes,” Phys. Rev. Lett., vol. 43, no. 1566,
1979.
73
[51] S. Weinberg, “Varieties of baryon and lepton nonconservation,” Phys. Rev. D, vol. 22, no. 1694,
1980.
[52] P. Minkowski, “µ→ eγ at a rate of one out of 1-billion muon decays?,” Phys. Lett. B, vol. 67, no. 421,
1977.
[53] R. N. Mohapatra and G. Senjanovic, “Neutrino mass and spontaneous parity violation,” Phys. Rev.
Lett., vol. 44, no. 912, 1980.
[54] R. N. Mohapatra and G. Senjanovic, “Neutrino masses and mixings in gauge models with sponta-
neous parity violation,” Phys. Rev. D, vol. 23, no. 165, 1981.
[55] J. Schechter and J. W. F. Valle, “Neutrino masses in su(2) x u(1) theories,” Phys. Rev. D, vol. 22,
no. 2227, 1980.
[56] G. Lazarides, Q. Shafi, and C. Wetterich, “Proton Lifetime and Fermion Masses in an SO(10)
Model,” Nucl.Phys., vol. B181, pp. 287–300, 1981.
[57] T. P. Cheng and L.-F. Li, “Neutrino masses, mixings, and oscillations in su(2)xu(1) models of elec-
troweak interactions,” Phys. Rev. D, vol. 22, no. 2860, 1980. arXiv:1207.7235[hep-ex].
[58] W. Konetschny and W. Kummer, “Nonconservation of total lepton number with scalar bosons,” Phys.
Lett. B, vol. 70, no. 433, 1977. arXiv:1207.7235[hep-ex].
[59] R. Foot, H. Lew, X. He, and G. Joshi, “Renormalizable adjoint su(5),” Z. Phys. C, vol. 44, no. 441,
1989.
[60] E. Ma, “Pathways to naturally small neutrino masses,” Phys. Rev. Lett., vol. 81, no. 1171-1174,
1998. hep-ph/9805219.
[61] A. Zee, “A theory of lepton number violation, neutrino majorana mass, and oscillation,” Phys. Lett.
B, vol. 93, no. 389, 1980.
[62] L. Wolfenstein, “A theoretical pattern for neutrino oscillations,” Nucl. Phys. B, vol. 175, no. 93, 1980.
[63] E. Ma, “Verifiable radiative seesaw mechanism of neutrino mass and dark matter,” Phys. Rev. D,
vol. 73, no. 077301, 2006. hep-ph/0601225.
[64] S. Blanchet and P. F. Perez, “Baryogenesis via leptogenesis in adjoint su(5),” JCAP, vol. 0808,
no. 037, 2008. arXiv:0807.3740.
[65] K. Kannike and D. V. Zhuridov, “New solution for neutrino masses and leptogenesis in adjoint su(5),”
JHEP, vol. 07, no. 102, 2011. arXiv:1105.4546.
[66] P. Nath and P. F. Perez, “Proton stability in grand unified theories, in strings and in branes,” Phys.
Rept., vol. 441, no. 191-317, 2007. arXiv:1207.7235[hep-ex].
[67] J. Hewett, H. Weerts, R. Brock, J. Butler, B. Casey, et al., “Fundamental Physics at the Intensity
Frontier,” 2012.
74
[68] D. Emmanuel-Costa, C. Simoes, and M. Tortola, “The minimal adjoint-SU(5) x Z4 GUT model,”
JHEP, vol. 1310, p. 054, 2013.
[69] H. Georgi and S. L. Glashow, “Gauge theories without anomalies,” Phys.Rev. D, vol. 6, no. 429,
1972.
[70] L. Krauss and F. Wilczek, “Discrete gauge symmetry in continuum theories,” Phys. Rev. Lett.,
vol. 62, no. 1221, 1989.
[71] L. E. Ibanez and G. G. Ross, “Discrete gauge symmetry anomalies,” Phys. Lett. B, vol. 260, no. 3,4,
1991.
[72] I. Dorsner, S. Fajfer, and N. Kosnik, “Heavy and light scalar leptoquarks in proton decay,” Phys. Rev.
D, vol. 86, no. 015013, 2012. arXiv:1205.2671.
[73] S. Weinberg, “Approximate symmetries and pseudogoldstone bosons,” Phys. Rev. Lett., vol. 29,
no. 1698-1701, 1972.
[74] P. O. Ludl and W. Grimus, “A complete survey of texture zeros in the lepton mass matrices,” 2014.
arXiv:1406.3546.
[75] J. Schechter and J. W. F. Valle, “Neutrinoless double beta decay in su(2)xu(1) theories,” Phys. Rev.
D, vol. 25, no. 2951, 1982.
75
Appendix A
Renormalization Group Equations and
complementary data
In the MS-bar renormalization scheme, the RGEs for gauge couplings gi are given, at 1-loop level,
by [31]dgi
d(lnµ)=
1
16π2big
3i , (A.1)
where µ refers to the energy scale and the bi are one-loop beta coefficients accounting for contributions
to gauge coupling running from the effective fields at the energy range of interest. Using αi = g2i /(16π2)
(fine structure notation), the solutions to (A.1) are
α−1i (µ2) = α−1
i (µ1)− bi2πln(
µ2
µ1), (A.2)
where µ2 > µ1. We can obtain the bi coefficients through
bi =1
3
∑R
s(R)ti(R)∏j 6=i
dj(R), (A.3)
where R is a field in a given representation, ti(R) is the Dynkin corresponding to that representation,
di(R) represents its dimension and s(R) is 1/2 for real scalars, 1 for complex scalars, 2 for Weyl fermions,
4 for Dirac fermions and -11 for gauge bosons.
For SU(2) we can get ti(R) and di(R) from the number of boxes in a Young diagram, a1, using
d2(R) = a1 + 1, (A.4)
t2(R) =a1(a1 + 1)(a1 + 2)
12. (A.5)
For SU(3) we follow a similar procedure to arrive at
d3(R) =(S + 2)(S + P + 1)
2, (A.6)
77
t3(R) =(S + 2)(S + P + 1) [S(S + 3)− P ]
48, (A.7)
where S ≡ a1 + a2 and P ≡ a1a2, a1 and a2 are the number of columns with one and two boxes,
respectively. It can be shown that these diophantine equations have the following solutions:
S = ui − 2
P =2d3(R)
ui− ui + 1,
(A.8)
with ui being the integer divisors of 2d3(R) within
2 3√d3(R) ≤ ui ≤
1 +√
1 + 8d3(R)
2. (A.9)
As for the U(1) case, this group is abelian so its representations are unidimensional, while t1(R) = 35YR,
where 35 accounts for the SU(5) hypercharge normalization and YR is the SM hypercharge.
Using these results we compute, with the SM particle content,
b1 =3
5
[ng(
2
3(1
2)2 × 2 +
2
3× (1)2 +
2
3(1
6)2 × 2× 3 +
2
3(2
3)2 × 3 +
2
3(1
3)2 × 3) +
1
3(1
2)2 × 2
]=
41
10,
(A.10)
b2 = −11
3× 2 +
1
3
1
2+ ng
1
3(1
2× 2× 3 +
1
2× 2) = −19
6, (A.11)
b3 =11
3× 3 + ng(
2
3
1
2× 2
2
3
1
2+
2
3
1
2) = −7. (A.12)
78
Appendix B
Renormalization Group Equations and
complementary data
B.1 Generalized Gell-Mann Matrices
With possible exception of a normalization factor, the generators of any SU(n) group are given, in the
fundamental representation, by generalized Gell-Mann matrices. We will display these matrices for the
SU(5) case and, since GSM is a maximal subgroup of this special unitary group, we will be able to match
some of these generators to the SM’s SU(3) and SU(2).
SU(3) Generators:
λ1 =
0 1 0
1 0 0
0 0 0
λ2 =
0 −i 0
i 0 0
0 0 0
λ3 =
1 0 0
0 −1 0
0 0 0
λ4 =
0 0 1
0 0 0
1 0 0
λ5 =
0 0 −i
0 0 0
i 0 0
λ6 =
0 0 0
0 0 1
0 1 0
λ7 =
0 0 0
0 0 −i
0 i 0
λ8 =
1√3
1 0 0
0 1 0
0 0 −2
.
(B.1)
79
Mixed Quantum Numbers Generators:
λ9 =
1 0
0 0
0 0
1 0 0
0 0 0
λ10 =
−i 0
0 0
0 0
i 0 0
0 0 0
λ11 =
0 0
1 0
0 0
0 1 0
0 0 0
λ12 =
0 0
−i 0
0 0
0 i 0
0 0 0
λ13 =
0 0
0 0
1 0
0 0 1
0 0 0
λ14 =
0 0
0 0
−i 0
0 0 i
0 0 0
λ15 =
0 1
0 0
0 0
0 0 0
1 0 0
λ16 =
0 −i
0 0
0 0
0 0 0
i 0 0
λ17 =
0 0
0 1
0 0
0 0 0
0 1 0
λ18 =
0 0
0 −i
0 0
0 0 0
0 i 0
λ19 =
0 0
0 0
0 1
0 0 0
0 0 1
λ20 =
0 0
0 0
0 −i
0 0 0
0 0 i
.
(B.2)
SU(2) Generators:
λ21 =
0 1
1 0
λ22 =
0 −i
i 0
λ23 =
1 0
0 −1
. (B.3)
Diagonal (Hypercharge) Generator:
λ24 =1√15
2
2
2
−3
−3
. (B.4)
This generator is particularly important in SU(5) GUTs, as it is closely related to electric charge quan-
80
tization. In fact, it has a different eigenvalue for the SU(3) and SU(2) Cartan subalgebras and, after
being multiplied by a normalization factor, it returns the field’s SM hypercharges, so, considering the
definition of electric charge in Chapter 1, charge quantization is achieved (recall that hypercharges were
not determined by gauge group factors in the SM).
B.2 Matter and Higgs Representations
In this section we use the following notation: greek indexes correspond to SU(5), so α, β, γ = 1, ..., 5, the
a, b, c, d= 1, 2, 3 are SU(3) indexes, the r, s, t= 4, 5 are SU(2) indexes, σI represent the Pauli matrices
and λA represent Gell-Mann matrices.
In terms of GSM quantum numbers, the fermionic fields in the Adjoint SU(5) model decompose as
5∗ = dC ⊕ L,
10 = Q⊕ uC ⊕ eC ,
24 = ρ8 ⊕ ρ3 ⊕ ρ(3,2)− 12
⊕ ρ(3∗,2) 12
⊕ ρ0.
(B.5)
Considering this, we may write
5∗a = (dC)a, 5∗r = εrsL2, (B.6)
10ab =1√2εabc(uC)c, 10ar =
1√2Qar, 10rs =
1√2εrseC . (B.7)
Taking into account the fact that
ραβ = ρβ∗α ,∑α=1
5ραα = 0, (B.8)
we get
ρab =1
2ρA8[λA]ab
+
√1
15δab ρ0,
rhoar =[ρ(3,2)
]ar,
ρrs =1
2ρI3[σI]rs−√
3
20δrsρ0.
(B.9)
Proceeding in the same way for the model’s scalars we obtain
5H = T1 ⊕H1,
24H = Σ8 ⊕ Σ3 ⊕ Σ(3,2)− 12
⊕ Σ(3∗,2) 12
⊕ Σ0,
45H = S(8,2) 310
⊕ S(6∗,1)− 15
⊕ S(3∗,2)− 710
⊕ S(3∗,1) 45
⊕∆⊕ T2 ⊕H2,
(B.10)
where ∆ ∼ (3, 3,−1/5). The Hs and Ts are Higgs doublets and coloured triplets, respectively. As for the
81
45H , we decompose it as
45abc = εabd[S(6∗,1)
]dc
+ δacTb2 − δbcT a2 ,
45abr = εabc[S(3∗,2)
]cr,
45arb =1
2SAr8
[λA]ab
+ δabHr2 ,
45ars = ∆ars + δrsT
a2 ,
45rsa = εrs[S(3∗,1)
]a,
45rst = −3(δrtHs2 − δstHr
2 ),
(B.11)
with[S(6∗,1)
]ab
=[S(6∗,1)
]ba
. The ∆ fields may be decomposed as
∆ ≡ ∆I σI
2=
1
2
∆−1/3√
2∆2/3
√2∆−4/3 −∆−1/3
. (B.12)
B.3 Scalar Potential
When one considers the existence of a 5H , a 24H and a 45H scalar representations, the most general
and renormalizable scalar potential may be written as
V (5H , 24H , 45H) = V1(5H)+V2(24H)+V3(45H)+V4(24H , 45H)+V5(5H , 24H)+V6(5H , 45H)+V7(5H , 24H , 45H),
(B.13)
where
V1(5H) = −µ25
25†H5H +
λ1
4(5†H5H)2, (B.14)
V2(24H) =− µ224
2Tr(242H
)+λ2
2Tr(242H
)2+a1
3Tr(243H
)+
+λ3
2Tr(244H
),
(B.15)
V3(45H) =− µ245
245αβHγ45∗γHαβ + λ4(45αβHγ45∗γHαβ)2+
+ λ545αβHγ45∗δHαβ45κλHδ45∗γHκλ + λ645αβHγ45∗δHαβ45κγHλ45∗λHκδ+
+ λ745αδHβ45∗βHαγ45κγHλ45∗λHκδ + λ845αγHδ45∗βHγε45κδHα45∗εHκβ+
+ λ945αγHδ45∗βHγε45κεHα45∗δHκβ + λ1045αγHδ45∗βHγε45κδHβ45∗εHκα+
+ λ1145αγHδ45∗βHγε45κεHβ45∗δHκα,
(B.16)
V4(24H , 45H) =a245αβHγ24γHδ45∗δHαβ + λ1245αβHγ45∗γHαβ24δHε24εHδ+
+ λ1345αβHγ24γHδ24εHβ45∗γHδε + λ1445αβHγ24γHβ24δHε45∗εαδ+
+ λ1545αβHγ24γHε24δHβ45∗εHαβ + λ1645αβHγ24κHα24λHκ45∗γHλβ+
λ1745αβHγ24γHκ24κHλ45∗λHαβ ,
(B.17)
V5(5H , 24H) = a35†H24H5H + λ185†H5H Tr{242H}+ λ195†H242
H5H , (B.18)
V6(5H , 45H) = λ2045αβHγ45∗γHαβ5∗Hδ5δH + λ2145αβHδ5
∗Hγ45∗γHαβ5δH + λ2245αβHγ45∗γHαδ5
∗Hβ5δH , (B.19)
82
and
V7(5H , 24H , 45H) = a45∗Hα45αβHδ24δHβ + λ235∗Hα45αβHδ24γHbeta24δHγ +H.c., (B.20)
where the λis and aj are constants.
83
Appendix C
Discrete Symmetries that suppress
Proton Decay and resulting Mass
Matrices
In this Appendix we present the different solutions (as far as the Mu and Md mass matrices are con-
cerned) with no tree-level scalar-mediated proton decay, that may be achieved for specific discrete sym-
metries. We also show the case where the new symmetry is continuous (U(1)) to make it clear that some
solutions can only be achieved through discrete symmetries. The row where a given solution is situated
indicates the number of zeros in Mu and the column indicates the number of zeros in Md. Elements
can be null if there is no charge assignment for which there is a solution and Mu, Md have the specified
amount of zeros or a non-null number that represents the possible different solutions with the specified
amount of mass matrix zeros where the textures and the symmetries displayed by the Yukawa matrices
at GUT scale are taken into account.
Z2
0 1 2 3 4 5 6
0 4 0 0 0 0 0 0
1 0 0 0 0 0 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
Z3
85
0 1 2 3 4 5 6
0 0 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 42 0 0 0
4 2 6 12 12 12 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 6
Z4
0 1 2 3 4 5 6
0 4 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 42 138 54 0
5 0 0 0 30 105 78 0
6 0 0 0 0 0 0 6
Z5
0 1 2 3 4 5 6
0 0 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 42 111 54 0
5 0 0 0 0 27 84 42
6 0 0 0 0 24 96 42
C.1 Results for ZN × ZM symmetries
Z2 × Z2
0 1 2 3 4 5 6
0 4 0 0 0 0 0 0
1 0 0 0 0 0 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 27 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
86
Z2 × Z3
0 1 2 3 4 5 6
0 4 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 42 0 0 0
4 2 6 12 54 123 54 0
5 0 0 0 0 54 168 84
6 0 0 0 0 45 174 108
Z2 × Z4
0 1 2 3 4 5 6
0 4 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 42 153 54 0
5 0 0 0 30 105 120 42
6 0 0 0 0 0 0 126
Z3 × Z3
0 1 2 3 4 5 6
0 0 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 42 0 0 0
4 2 6 12 12 12 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 78
C.2 Results for continuous symmetries
-5 to 5
0 1 2 3 4 5 6
0 0 0 0 0 0 0 0
1 4 6 12 0 3 0 0
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 42 111 54 0
5 0 0 0 0 27 84 42
6 0 0 0 0 24 96 78
87