Astrophysical implications of the Bumblebee model of
Spontaneous Lorentz Symmetry Breaking
Gonçalo Dias Pereira Guiomar
Thesis to obtain the Master of Science Degree in
Engineering Physics
Supervisor(s): Prof. Dr. Vítor Manuel dos Santos Cardoso
Prof. Dr. Jorge Tiago Almeida Páramos
Examination Committee
Chairperson: Prof. Dr. Ana Maria Vergueiro Monteiro Cidade Mourão
Supervisor: Prof. Dr. Jorge Tiago Almeida Páramos
Members of the Committee: Prof. Dr. Amaro José Rica da Silva
November 2014
Acknowledgments
This thesis could not have been possible without the help and guidance of my supervisor, Professor
Jorge Páramos, with whom I rediscovered the joy of doing Physics. His immense patience when dealing
with my incompetence, along with his immense knowledge of unorthodox working places made this
work a fun and rewarding experience. Also, I would like to thank Professor Vítor Cardoso for his help
and availability in the process of realizing this thesis.
For my family, I am truly grateful for your continuous support and for providing me the opportunity
of realizing my goals, no matter how uncertain they might have seem in the past.
For my friends, who accompanied me throughout this journey, thank you for joining me in my
culinary digressions.
This last paragraph I dedicate to Geisa, for helping me collapse onto a better state of being.
i
Abstract
In this work the Bumblebee model for spontaneous Lorentz symmetry breaking is considered in the
context of spherically symmetric astrophysical bodies. A discussion of the modied equations of motion
is presented and constraints on the parameters of the model are perturbatively obtained. Along with
this, a detailed review of this model is given, ranging from the questioning of the basic assumptions of
General Relativity, to the role of symmetries in Physics and the Dark Matter problem.
Keywords
General Relativity, Bumblebee Model, Lorentz Symmetry Breaking, Stellar Equilibrium (English)
iii
Resumo
Neste trabalho, consideramos o modelo Bumblebee para a quebra espontânea da simetria de Lorentz
no contexto de corpos celestes com simetria esférica. Uma discussão das equações de movimento mod-
icadas é apresentada, juntamente com os constragimentos do modelo obtidos de modo perturbativo.
De modo a contextualizar o modelo e o problema em questão, uma revisão é apresentada, onde se
abordam temas tais como os fundamentos da Relatividade Geral, o papel das simetrias na física e o
problema da matéria escura.
Palavras Chave
Relatividade Geral, Modelo Bumblebee, Quebra da simetria de Lorentz, Equilíbrio Estelar
v
Contents
1 Introduction 2
1.1 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Essential concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 General Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 The limits of Einstein's Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2 Astrophysics 9
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Tolman-Openheimer-Volkov equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Polytropes and the Lane-Emden Equation . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3 Symmetry Breaking 15
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 Explicit and Spontaneous Breaking of Symmetries . . . . . . . . . . . . . . . . . . . . . 17
3.3 Observer and Particle LSB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Standard Model Extension and LSB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4 Vector Theories 25
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Aether Theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3 The Bumblebee Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5 Astrophysical constraints on the Bumblebee 37
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2 Static, spherically symmetric scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.3 Perturbative Eect of the Bumblebee Field . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.4 Numerical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6 Conclusions 45
Bibliography 47
vii
List of Figures
2.1 Lane-Emden solution for a spherical body in hydrostatic equilibrium. . . . . . . . . . . . 14
3.1 Snowakes generated through a Linenmayer rule set for the rst, second and third
iterations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 The Mexican Hat potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1 The Bullet Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.1 Prole of the relative perturbations pb/p0, pV /p0, ρb/ρ0 and ρV /ρ0 induced by the
Bumblebee. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.2 Allowed region (in grey) for a relative perturbation of less than 1% for pb and ρb. . . . . 43
5.3 Allowed region (in grey) for a relative perturbation of less than 1% for pV and ρV . . . . 44
ix
List of Tables
5.1 Selected stars that used as models for the numerical analysis of the Bumblebee pertur-
bation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
5.2 Table of non-dimensional parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
xi
Abbreviations
GR - General Relativity
TOV - Tolman-Oppenheimer-Volko
LE - Lane-Emden
EP - Equivalence Principle
LLI - Local Lorentz Invariance
LPI - Local Position Invariance
EP - Equivalence Principle
WEP - Weak Equivalence Principle
SEP - Strong Equivalence Principle
LSB - Lorentz Symmetry Breaking
PPN - Parametrized Post Newtonian
CPT - Charge Parity Time Symmetry
GZK - Greisen-Zatsepin-Kuzmin
CMBR - Cosmic Microwave Background Radiation
CMBR - Cosmic Microwave Background Radiation
HI-RES - High Resolution Fly's Eye
SME - Standard Model Extension
SM - Standard Model
SSB - Spontaneous Symmetry Breaking
QED - Quantum Electrodynamics
MOND - Modied Newtonian Dynamics
FLRW - Friedmann-Lemaitre-Roberston-Walker
VEV - Vacuum Expectation Value
1
1Introduction
Contents
1.1 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Essential concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 General Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 The limits of Einstein's Relativity . . . . . . . . . . . . . . . . . . . . . . . 5
2
1.1 Thesis Outline
1.1 Thesis Outline
This work will begin by a small introduction of the relevant concepts needed from General Relativity
in the subsection 1.2 and the current observed limitations of the theory in subsection 1.4. The Tolman-
Oppenheimer-Volko equations are then introduced in Chapter 2 along with the Lane-Emden (LE)
model for spherical bodies, which will be needed in order to understand Chapter 5. In Chapter 3, the
concept of Lorentz Symmetry Breaking (LSB), in both its explicit and spontaneous form, is introduced
as a way to portray one of the fundamental properties of the model being used, the Bumblebee Model.
This will later be introduced in 4 as a particular case of the more general Aether Theories. The two
nal chapters (5,6) contain the kernel of this work: The application of the models presented as a way
of constraining their parameters using a spherical astrophysical body modeled by a polytrope. This
thesis closely follows the work done in [1].
1.2 Essential concepts
Presently, Einstein's general theory of relativity serves as a tool to understand a wide range of
phenomena. From the dynamics of compact astrophysical bodies, such as stars and black holes, to
cosmology, its striking predictions have sustained this theoretical framework and made it one of the
most relevant scientic achievements in the history of science [2].
The experimental conrmation of the existence of gravitational lensing, time dilation and gravi-
tational redshift have solidied this fact but this by no means imply that the theory is completely
correct; in fact, such consistence is a motivating factor for testing its limits even further.
Questioning the basic assumptions of General Relativity is thus a valid way to achieve this goal, as
they are the fundamental rule set from which it emerges. These basic assumptions can be expressed
in the following manner [3]:
• Weak Equivalence Principle - Bodies in free fall have the same acceleration independently of
their compositions.
• Local position invariance - The rate at which a clock ticks is independent of its position.
• Local Lorentz invariance - The rate at which a clock ticks is independent of its velocity.
This work shall focus on the second and third basic assumptions presented above. A more detailed
discussion will be presented in 3 in how this principle can be used to test the validity of physical
theories.
As a short introduction to the review that will be made in the following section, a small appetizer
is given in the context of special relativity, showing how one can test the limits of such theory.
The tests are based on testing the principles of relativity (physical laws are independent of the
inertial frame of reference used to describe them) and the constancy of the speed of light. In the case
of inertial frames of reference (which can always be found, at least in the vicinity of any given point in
spacetime), this translates into invariance under Lorentz transformations, a tenet of Special Relativity:
deviations from these transformations would also imply deviations from the underlying principles.
3
1. Introduction
This can be approached through the Robertson-Sexl-Mansouri formalism, which consists on the
known Lorentz transformations,
t′ =t− vx
c2√1− v2
c2
, x′ =x− vt√1− v2
c2
, y′ = y, z′ = z, (1.1)
altered in a way as to express a preferred frame of reference Σ(T, ~X). The transformations now take
the form,
T =t− ~ε.~xa
, ~X =~x
d−(
1
d− 1
b
)~v~x
v2~v + ~vT. (1.2)
With these transformations, it can be shown [3] that through an expansion of a, b, d and ~ε around
v/c2 one can obtain the following expression for the relative shift in the two way speed of light,
c2(θ, v)
c2(0, v)−1 = sin2 θ
[(δ−β)
(v
c
)2
+
(3δ2 − β2
4−β2−
β
2(1 + δ)− δ2 +
3
4(β− δ)2 cos 2θ
)(v
c
)4], (1.3)
with θ the angle between the velocity ~v of the frame of reference and the path of light, and αn, βn, δn, εn
the expansion terms of a(v), b(v), d(v) and ε respectively with n = 1, 2 the order of the expansion.
If any of the referred expansion terms are veried to be non-vanishing in some measurement, this
would imply the violation of Lorentz invariance.
1.3 General Relativity
Having introduced the general method, we now move onto dene the basic formalism of General
Relativity, along with the conventions that will be used in the remainder of this work.
General Relativity proceeds by claiming that the principle of relativity indeed applies to all frames
of reference, and not just inertial ones. Thus, instead of considering only invariance under Lorentz
transformations, it imposes general (dieomorphism) invariance, i.e. the laws of physics are invariant
under general (dierentiable) coordinate transformations.
The need for having a priori coordinate invariance implies that we need to use a scalar Lagrangian
density which, depends on the elds and their derivatives (up to second order in order to avoid Ostro-
gradsky's instabilities [6]). In Riemannian spacetime, this eld is the metric tensor and its derivatives
are embodied in the Ricci curvature scalar, leading to the standard Einstein-Hilbert action L = R. If,
however, we promote R to a fundamental variable, then more general forms are admissible
L =√|g| [Λ + bR+ c∇µ∇νRµν + ...+ f(R)] , (1.4)
with f(R) a possible function of the metric from which we can build modied versions of the basic
Einstein-Hilbert (which is obtained by selecting the linear terms alone).Variation of the Einstein-Hilbert
action leads to
δS =
∫dx4√−g
[− 1
16πG
(Rµν − 1
2gµνR
)+
1
2Tµν
]δgµν , (1.5)
with
Tµν = − 2√−gδ(√−gL)
δgµν. (1.6)
4
1.4 The limits of Einstein's Relativity
Imposing a null variation yields the Einstein's equations of motion,
Gµν ≡ Rµν − 1
2gµνR = 8πGTµν (1.7)
where G is the gravitational constant. This identity tells us one of the most important results of this
theory: The distribution of energy in spacetime dictates its curvature.
The doubly contracted Bianchi identities ∇νGµν = 0 then imply the conservation of the energy-
momentum tensor ∇νTµν = 0. In the absence of matter, one would obtain R = 0 for the Ricci scalar;
in a static and spherically symmetric spacetime, this leads to the Schwarzschild solution.
The following chapters will use these results to derive the Tolman-Oppenheimer-Volko equation
in chapter 2 or the results presented in the nal chapter.
1.4 The limits of Einstein's Relativity
Having introduced the basic formalism that will be used throughout this work, we shall now discuss
some of the known experimental limits that constrain General Relativity as well as the methods used
in obtaining them. These observational bounds can be used both to test the validity of the foundations
of GR and alternative theories of gravity [2].
Varying the action S =∫ √−gµνdxµdxν one then gets the geodesic equation of motion
d2xµ
dτ2= Γµαβ
dxα
dτ
dxβ
dτ. (1.8)
In the static weak-eld limit, the metric can be written as
g0i = gi0 = 0, g00 = −1− 2ΦNc2
, gij =
(1 +
2ΦNc2
)δij (1.9)
with ΦN the Newtonian gravitational potential. Inserting this metric into the geodesic equation we
get Newton's second law of motion for the (0, 0) component
d2xi
dt2= −Γi00 = −∂ΦN
dxi(1.10)
and, from Einstein's equations of motion 1.7, we get the Poisson equation
∇2ΦN = 4πGρ. (1.11)
This simply means that the metric plays the role of the Newtonian gravitational potential, with the
Christoel symbol behaving analogously to an acceleration.
Thus, by changing the metric one could explore how certain aspects of a theory cascade into the
equations of motion of the system.
1.4.1 Parametrized Post Newtonian Formalism
We now introduce the Parametrized Post Newtonian Formalism, or PPN for short. Behind every
metric theory of gravity lies the underlying principle that the eponymous tensor directly aects the way
in which the gravitational eld interacts with matter. As such, this formalism serves as a generalization
of the metric to include parameters which express certain symmetries and laws of invariance of the
5
1. Introduction
system, which then allows us to measure the deviation of these GR parameters in relation to Newtonian
gravity.
Assuming Local Lorentz and Position Invariance, along with conservation of momentum, it can be
shown that the metric tensor in PPN is given by,
g00 = −1 + 2U − 2βU2 − 2ξΦW + (2γ + 2α3 + ζ1 − 2ξ)Φ1 + 2(3γ − 2β + 1 + ζ2 + ξ)Φ2 (1.12)
+2(1 + ζ3)Φ3 + 2(3γ + 3ζ4 − 2ξ)Φ4 − (ζ1 − 2ξ)A− (α1 − α2 − α3)ω2U
−α2ωiωjUij + 2(2α3 − α1)ωiVi +O(ε3)
g0i = −1
2(4γ + 3 + α1 − α2 + ζ1 − 2ξ)Vi −
1
2(1 + α2 − ζ1 + 2ξ)Wi (1.13)
−1
2(α1 − 2α2)ωiU − α2ω
jUij +O(ε5/2)
gij = (1 + 2γU)δij +O(ε2), (1.14)
where β is a measure of the non-linearity of the law of superposition of gravitational elds, γ mea-
sures the curvature of the spacetime created per unit rest mass, α1, α2, α3 measure deviations from
Lorentz invariance, ζ the violation of Local Position invariance and α3, ζ1, ζ2, ζ3, ζ4 measure the pos-
sible violation of the conservation of momentum. The expressions for the gravitational potentials
U,Uij ,Φi,A, Vi,Wi are given in [3]. GR is characterized by β = γ = 1, and all other PPN parameters
vanish.
Current experimental tests on these parameters, particularly the couple γ, β, are the result of
measurements made by the Cassini 2003 spacecraft and helioseismology show that γ − 1 ≈ 2.3× 10−5
and β − 1 ≈ 3× 10−4, respectively [7].
1.4.2 Equivalence Principle, Lorentz and Position Invariance
We now move to the testing of the postulates presented in the previous section; the equivalence
principle and both the Local invariance principles (position and Lorentz), which are the fundamental
groundwork from which GR is built from. Tests for the Equivalence Principle can be divided into two
groups: those that test the weak version (WEP) and those that test the strong equivalence principle
(SEP).
The WEP states that the Equivalence Principle (all non-gravitational laws should behave in free-
falling frames as if there was no gravity) is satised by all interactions except that of gravity. One
could test its validity by simply measuring the dierence in the free-fall accelerations between two test
bodies a1 and a2 [3],
∆a
a=
2(a1 − a2)
a1 + a2=
(MG
MI
)
1
−(MG
MI
)
2
= ∆
(MG
MI
). (1.15)
where MG and MI represent the gravitational and inertial masses, or by directly measuring the ratio
MG/MI . For the latter, various experiments have been made where the most recent and strongest
constraint being given by Adelberger in Ref. [8] of |1−MG/MI | ≈ 1.4× 10−13.
Another consequence of the WEP is the existence of a gravitational Doppler eect in bodies which
travel trough a changing gravitational potential. One of the most historically relevant experiments
6
1.4 The limits of Einstein's Relativity
was made by Robert Pound and Glen Rebka [9], where photons where emitted by a moving source
and later absorbed by a stationary target located at the top of a tower. The absorption was only
possible as the relativistic doppler shift of the moving source cancelled the gravitational doppler eect
of the graviational eld of the earth. The measured change in frequency ∆ν/ν = 2.57± 0.26× 10−15
conrmed the existence of this eect, as predicted by GR.
The SEP, on the other hand, states that every measurement is independent of the velocity and
position of the laboratory, even accounting for the self-energy of massive bodies such as stars and
black holes.
Testing the SEP implies measuring the contributions of this gravitational energy which were not
considered in the WEP. In order to accomplish this we resort to the PPN formalism introduced above,
introducing the quantity
∆
(MG
MI
)= η
(Ω
Mc2
)(1.16)
where Mc2 is the total mass energy of the body and Ω its negative gravitational self-energy. Here
η = 4β − γ − 3 is a combination of PPN parameters. This combination implies that in GR one should
have η = 0, since γ = β = 1.
We now move on to the two nal assumptions presented in the previous section: Local Lorentz and
Position Invariance, or LLI and LPI respectively.
The phenomenological eect of moving in a reference frame relatively to a stationary one can be
probed by assuming a cosmological vector eld which collapses onto a non-vanishing minimum via
spontaneous symmetry breaking. These types of models have been proposed by Kostelecký [10] and
its impact on solar system observables was discussed in [11]. The mechanism underlying the breaking
of the symmetry is explained in chapter 3.
Experimental searches for the breaking of this symmetry have been made through diverse physical
phenomena. As a violation of this symmetry could imply the breaking of the CPT symmetry, numerous
proposals have been made for the possible testing of this possibility [12].
As referred above, the PPN formalism is also a good way in which to infer the departure of an
experimental phenomenon from what would be expected from GR. In what regards the parameters
that relate to preferred frame eects, the more relevant of these is α2, for which the observational limit
is |α2| < 4 × 10−7[13] and reects the existence of spin precession anomalies, along with α3, which
reects self-acceleration eects. For the latter, a measurement was made via pulsar statistics in order
to measure its deviation if it is non-vanishing, yielding the constraint α3 < 2.2× 1020 [14].
Another way to measure possible Lorentz violation is through the study of the Greisen and Zatsepin
& Kuzmin (GZK) cut-o. This stems from the interaction between protons with energies of the order
of 1020eV and Cosmic Microwave Background Radiation (CMBR) photons in nuclear reactions of the
type
p+ γCMB → p+ π0. (1.17)
Due to these reactions, the primary protons would have their energy decreased, suering a type of
7
1. Introduction
friction from the cosmic background radiation. The threshold for these reactions is given by
Ef =m2π + 2mπmp
4EγCMB≈ 1019eV, (1.18)
which implies that above this energy value, we should not observe cosmic rays with these energies on
Earth.
Conrmation of the GZK limit through experiments such as the Pierre Auger Observatory [15]
would impose strong constraints [16] on the possible observation of LV in the QED as discussed above.
As discussed in Ref. [17], the detection of those 1019eV photons would imply,
|a1| ≤ 10−25, |a2| = |a3| ≤ 10−7 (1.19)
which would indicate a very weak presence of LV at the high energy scale. However, recent results from
both HI-RES (High Resolution Fly's Eye) and the Pierre Auger collaborations have shown evidence
that the cut-o has not been statistically broken, which implies that Lorentz invariance has also been
maintained [18].
In what regards LPI, the already mentioned Pound-Rebka experiment can serve as a test for this
eect, with the measurement of the change in frequency given by
∆ν
ν=
(1 + µ)U
c2(1.20)
were µ = 0 in GR, would give an idea of how far the assumption holds. Measurements made with
hydrogen-maser frequencies on earth and on altitudes of ten thousand kilometres yield µ of |µ| <2× 10−4 [19].
Although LI has these strong constraints, the fact is that there is still room for obtaining deviations
in energy scales that are unreachable today [13, 20]. It is through this window of opportunity that we
shall peer through in the following sections.
8
2Astrophysics
Contents
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Tolman-Openheimer-Volkov equation . . . . . . . . . . . . . . . . . . . . . 10
2.3 Polytropes and the Lane-Emden Equation . . . . . . . . . . . . . . . . . . 12
9
2. Astrophysics
2.1 Introduction
In this section, the hydrostatic equilibrium equation is obtained via an approximation of the known
Tolmann-Oppenheimer-Volko equation which, is derived directly from Einstein's equations of motion.
Assuming a polytropic equation of state, these equations are the so called Lane-Emden dierential
equations, which have a solution depending on the polytropic index n alone. This model will be later
used in chapter 5 as a description of a non-perturbed star from which the subsequent analysis of the
Bumblebee model will follow.
2.2 Tolman-Openheimer-Volkov equation
The description of a star's interior when in hydrostatic equilibrium can be obtained in General
Relativity through the choice of an appropriate energy-momentum tensor along with a static spherically
symmetric metric.
Assuming that the uid inside the star behaves like a perfect uid, then the appropriate energy-
momentum tensor is given by,
Tµν = ρuµuν + p(gµν + uµuν), (2.1)
where the signature (−,+,+,+) is used both in this case and throughout the remainder of this work.
Given the static, spherically symmetric geometry, the Birkho metric is chosen, as it is described
by the line element,
ds2 = −e2ν(r)dt2 + e2λ(r)dr2 + r2dθ2 + r2 sin2(θ)dφ2 (2.2)
and so the energy-momentum tensor is given by
Tµν = diag(ρe2ν(r), pe2λ(r), pr2, pr2 sin2(θ)) (2.3)
and the trace by
T = gµνTµν = −ρ+ 3p. (2.4)
Using the trace-reversed form of the Einstein eld equations referred in Chapter 1
Rµν − 8π
(Tµν −
1
2gµνT
)= 0 (2.5)
the equations gttRtt,grrRrr and g
θθRθθ are then, respectively:
e2(ν−λ)
r
[r(λ′ − ν′)ν′ − rν′′ − 2ν′
]= −4πGe2ν(3p+ ρ) (2.6)
1
r
[2λ′ + r(λ′ − ν′)ν′ − rν′′
]= −4πGe2λ(p− ρ) (2.7)
e−2λ
[1 + e2λ − rλ′ + rν′
]= −4πGr2(p− ρ) (2.8)
Subtracting 2.6 and 2.7 then gives,
− 2e−2λ(λ′ + ν′)
r= 8πG(p+ ρ). (2.9)
10
2.2 Tolman-Openheimer-Volkov equation
For that we need to nd a relation between the density and the metric function λ(r). By using the
combination,
gttRtt + grrRrr − 2gθθRθθ = 0 (2.10)
we nd that
2[1− 8πGρr2 + e−2λ(−1 + 2λ′r)]
r2= 0 (2.11)
which gives us, through some manipulation, the equation
1− e−2λ + 2rλ′e−2λ = 8πGρr2 (2.12)
d
dr
(r − re−2λ
)= 8πGρr2. (2.13)
We can now obtain a relation pertaining λ
e−2λ = 1− 2Gm
r(2.14)
which can be obtained by calculating the mass through
m(r) =
∫ r
0
4πρ(r′)r′2dr′. (2.15)
After dierentiating λ one gets
− 2λ′e−2λ = G2m′r − 2m
r2= 8πGρr − 2Gm
r2, (2.16)
which, by substitution into equation 2.9, gives us a relation between the remainder metric function
ν(r) and the pressure,
ν′ =Gm+ 4πGpr3
r(r − 2Gm). (2.17)
We know from the twice contracted Bianchi identities that
∇νGµν = 0, (2.18)
which implies that the conservation of the energy-momentum tensor ∇νTµν = 0, from the Einstein
eld equations presented in Chapter 1.
By selecting the radial component µ = r, we arrive at the conservation equation for a perfect and
static uid,
p′ + ν′(p+ ρ) = 0, (2.19)
which allows us to reach a dierential equation for the pressure inside the star:
dp
dr=
(G(4πGpr3 +Gm)(p+ ρ)
r2
)(1− 2Gm
r
)−1
(2.20)
which is the so called Tolmann-Oppenheimer-Volko equation for hydrostatic equilibrium.
In order to solve the set of Eqs. 2.15 2.19 2.20 , which consists of a system of 4 unknown functions
λ, ν, p, ρ and three independent equations, we need another equation that relates the pressure and the
11
2. Astrophysics
density, i.e. an equation of state p = p(ρ). This motivates the next section, where we will be describing
the polytrope equation of state and the Lane-Emden dierential equation for a star in hydrostatic
equilibrium, as well as an introduction to the methodology applied in Chapter 5 for obtaining the
equations for the unperturbed star.
In order to proceed into the next section, we will assume the conditions of hydrostatic equilibrium,
i.e.:
p(r) << ρ(r), 4πp(r)r3 << m(r), 2Gm(r) << r, (2.21)
then the equation 2.20 can be approximated by its hydrostatic equilibrium version,
dp
dr= −Gm
r2ρ (2.22)
dm
dr= 4πr2ρ (2.23)
which when combined give us the Poisson equation ∇φ = 4πGρ in spherical coordinates
1
r2
d
dr
(r2
ρ
dp
dr
)= −4πGρ. (2.24)
2.3 Polytropes and the Lane-Emden Equation
With the previous conditions in mind, we can now assume that the pressure is given by the poly-
tropic relation in equation 2.25 for a certain polytropic index n
p = Kρ1+1/n, (2.25)
where K is the polytropic constant and ρ0 the baryonic mass density. The polytropic index describes
the basic thermodynamical processes. Taking n = 1 we get an isobaric sphere, n = 0 an isometric
one and n = ∞ gives the isothermal condition for that same sphere. The adiabatic processes are
related by n = 1/(γ−1) within this framework, where γ represents the adiabatic coecient γ = cp/cV .
These indexes can provide crude approximations to known astrophysical bodies [21]. Degenerate star
cores found in giant gaseous planets can be studied using a polytropic index of n = 3/2, boundless
systems which were rst use in the description of stellar systems by Arthur Schuster in 1883 with n = 5
and the rst solar models, which were rst proposed by Arthur Eddington (known as the Eddington
standard model of stellar structure), circa 1916 for a polytropic index of n = 3. Although these models
are rather simplistic, they are still useful as rst order approximations, as they allow for an easier
algebraic manipulation.
Dierentiation of 2.25 gives
dp
dr=
(n+ 1
n
)Kρ
1ndρ
dr(2.26)
which, when inserted into equation 2.24 changes into
1
r2
d
dr
[n+ 1
n
r2
ρKρ
1ndρ
dr
]= −4πGρ. (2.27)
12
2.3 Polytropes and the Lane-Emden Equation
We can now rewrite this dierential equation in its dimensionless form, by selecting the transfor-
mations
ρ = ρcθn(χ) , p = pcθ
n+1(χ), (2.28)
The dierential equation is now given by
(n+ 1)Kρ
1+ 1n
c
4πGρ2c
1
r2
d
dr
(r2 dθ
dr
)= −θn. (2.29)
where the underscript c refers to the central values of the quantities.
We can now dene χ = r/rn, where rn is dened through the star's radius and the terms present
in equation 2.29,
R = rnχf =
[(n+ 1)K
4πG
] 12
ρ1−n2nc χf . (2.30)
where χf signals the boundary of the spherical body (θ(χf ) = p(χf ) = ρ(χf ) = 0).
With all these components, we can now assemble the dimensionless Lane-Emden dierential equa-
tion:
1
χ
d
dχ
(χ2 dθ
dχ
)= −θn. (2.31)
The analytical solution for this equation was found by Chandrasekar [22] for a set of polytropic
indexes n = (0, 1, 5). These are given by, respectively
θ(χ) = 1− 1
6χ2, θ(χ) =
sin(χ)
χ, θ(χ) =
1√1 + χ2
3
(2.32)
where the nal solution is innite in radial extent.
As we want to calculate the LE solution with n values around the Eddington solution with n = 3,
these exact solutions will not be useful. We thus proceed with the analysis of the system via a numerical
method. We know that the boundary conditions for this system are given by
θ(0) = 1 (2.33)
θ′(0) = 0 (2.34)
and so by making a series expansion of our equation at the point χ = 0, we can then calculate the
initial conditions of our system for some point χ = χi, and solve numerically the equation for any n
up to 5, where the analytical solution is radially innite as referred above.
Throughout this work, three values for the polytropic index where chosen, with the intent of showing
how its variation aects the model. The values are n = 2.8, 3, 3.2 and so the need to obtain a general
expression for the expansion of θ(χ) around χ = 0 arises.
Assuming that the solution θ is symmetric under the transformation χ→ −χ and that the solution
θ is analytic we can expand θ as,
θ(χ) ≈ 1− χ2
6+nχ4
120+
(5n− 8n2)χ6
15120+ ... (2.35)
13
2. Astrophysics
where the terms of order between n = 6 and n = nf where not shown for simplicity. This is useful in
a numerical computation where it might be necessary the computation of a non-zero initial value close
to the real null initial condition.
Using these conditions one then obtains the following proles for the Lane-Emden function θ(χ),
as can be seen in Figure 2.1
0 1 2 3 4 5 6 7 80
0.2
0.4
0.6
0.8
1
χ
θ(χ)
n = 2.9n = 3n = 3.1
Figure 2.1: Lane-Emden solution for the a spherical body in hydrostatic equilibrium. The roots of the threefunctions are χ2.8
f = 6.191, χ3f = 6.896 and χ3.2
f = 6.768.
Besides obtaining the prole θ(χ), we can also relate the stars fundamental observed properties, its
mass and radius, and relate them with the central density through the LE solution.
In order to go from the set (M,R) to the set (ρc, pc) from which the LE solution takes its scaling
parameters, we start by integrating the relation ρ0(r) = m′/4πr2 in order to obtain
M =
∫ R
0
4πρ(r)dr = 4πr3nρc
∫ χf
0
χ2θndχ. (2.36)
By inserting the LE equation into the integral one then gets,
M = 4π
[(n+ 1)K
4πG
]3/2
ρ3−n2nc
[−χ2
fθ′(χf )
](2.37)
which allows us to obtain K in terms of the mass and radius of the star and the polytropic index n
K = G
(4πn
n+ 1[−χ2
fθ′(χf )]
1−nn χ
n−3n
f
)(MR−1
)n−1n . (2.38)
The central pressure of the system is then given by inverting rn,
ρc = (4π)nn−1
[R
χf
(K(n+ 1)
G
)−1/2]−2nn−1
(2.39)
completing the parameters needed to completely describe the star.
14
3Symmetry Breaking
Contents
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2 Explicit and Spontaneous Breaking of Symmetries . . . . . . . . . . . . . 17
3.3 Observer and Particle LSB . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Standard Model Extension and LSB . . . . . . . . . . . . . . . . . . . . . . 21
15
3. Symmetry Breaking
3.1 Introduction
The concept of symmetry can be stated as an intrinsic property of a system which does not change
under certain transformations. Its usage throughout history can be traced back to both aesthetic and
technical means of describing apparent order and beauty, such as the apparent mirror symmetry of the
human body, or the more complex six-fold symmetry present in snow akes.
One of the most interesting aspects of this idea is the fact that when applied to physical systems,
the existence of symmetries restricts the possible outcomes that certain phenomena may have under
that system. This idea was stated by Pierre Curie in his Sur la symètrie dans les phénoménes physiques
[23], where he found that the thermal and electric properties of crystals varied with the underlying
crystal structure.
Figure 3.1: Snowakes generated through a Linenmayer rule set for the rst, second and third iterations.
A simple example can be given, as represented in gure 3.1. The snow ake was generated through
a formal rule set rst created by Aristid Lindenmayer, a Hungarian botanist whose work consisted in
trying to describe the evolution of the visual patterns present in the growth of yeast. The Linden-
mayer system takes an alphabet of symbols, an initial state for the system and a set of production
rules and applies the rules iteratively. In the case of our snowake, for each 60 angle a branch is
generated with two smaller branches, with each iteration adding onto the already growing pattern.
We have thus created a model that describes, up to a certain degree of certainty, a snowake which
one could nd in nature, only through the rules that describe its underlying symmetry. But what if
there were only two snowakes in the Universe, and one snowake was slightly dierent than the one
described by our model? Well, this would imply that one of the symmetries that described the "good"
snowake, would be broken by the "bad" one, and thus our model would have to be updated in order
to incorporate this fact.
This idea of looking for the underlying symmetries of a certain system and checking for their
existence in the Universe portrays the motivation behind the model used in this work, although it is
inserted in the larger context of both Einstein's Relativity and the Standard Model of Particle Physics.
The Principle of Relativity, as rst adopted by Einstein in his special theory of relativity, states
that the laws that govern the change of a physical system are independent of the chosen coordinate
systems, even when those are moving relatively to each other in uniform translational motion.
This principle, when coupled with the light postulate which guarantees a xed speed of light in all
16
3.2 Explicit and Spontaneous Breaking of Symmetries
inertial frames of reference, leads to the known Lorentz transformations between coordinate systems
which are moving relatively to each other.
This can be written succinctly, trough the matrix equation
x′ = Λ(v)x (3.1)
which, for a boost in the x direction for example, would be described by
ct′
x′
y′
z′
=
γ −βγ 0 0−βγ γ 0 0
0 0 1 00 0 0 1
c txyz
, (3.2)
representing thus the underlying symmetry of boosts in the x direction.
In General Relativity, this symmetry holds locally i.e. in distances small enough that the variations
in the gravitational eld are unnoticed. This is but a statement of the weak equivalence principle
which, as stated by Einstein, takes the form [24]
The outcome of any local non-gravitational experiment in a freely falling laboratory is in-
dependent of the velocity of the laboratory and its location in spacetime.
In the following sections, when speaking of Lorentz invariance, the concept behind the name will
follow closely what was described in the introduction. The goal is to provide a brief introduction to
the concept of Lorentz Symmetry Breaking (LSB) and its usage as a probe for studying fundamental
physical phenomena in both large and small scales.
3.2 Explicit and Spontaneous Breaking of Symmetries
The breaking of symmetries is a great source of richness in both physics and everyday life. Consider
then a hungry donkey who is placed between two stacks of hay and assume he always goes to the nearest
one. By the Principle of Sucient Reason put forward by Leibniz, the donkey cannot justify going
for either stack of hay and so it dies of hunger as it is unable to decide. Unless, of course, by some
unknown force of will, he breaks the symmetry of his dire situation and goes for one of those stacks
of hay! Thus it is justied that breaking a symmetry brings with it a lot more to life than simply
perturbing a perfect, albeit immutable situation.
In the case of physics, things are not so simple. We shall start by dening the two ways in which
these symmetries can be broken: explicitly and spontaneously, and provide some examples of these
kinds of phenomena.
For explicit symmetry breaking, the underlying symmetry that is broken appears in the description
of the physical laws themselves. This can take the form of certain terms in the Lagrangian density
that describes the system, as the dynamical equations are not invariant under that transformation.
One example of this is the parity violation in the weak interaction, which was initially proposed by
Yang & Lee circa 1950 [25] by studying the beta decay of cobalt-60. Up until this point the laws of
nature were assumed to be invariant under mirror reection transformations (as observed in gravity,
electromagnetism and the strong force), i.e. observing a certain experiment and its mirror reected
17
3. Symmetry Breaking
copy should yield the same results; however, in the case of beta decay, a preferred direction of the
emitted electrons was always observed.
In relativistic quantum mechanics, this symmetry portrays an invariance in the change between
particles and antiparticles within a system. Along with parity, charge conservation is also seems to be
broken in the weak interactions. The conjugation of both these symmetries (CP), was once thought
to be a valid symmetry of nature, but was later veried by Cronin & Fitch to be broken in the decay
of neutral kaons [26], which gave them the Nobel Prize in Physics in 1980.
Although the individual symmetries appear to not to be fundamental properties of our universe,
a combination of them does: a theorem proved by Schwinger, Pauli, Bell and Lüders around 1950
shows that the combination of charge conjugation, parity and time-reversal symmetries is conserved
in quantum eld theories which have Lorentz invariance, local causality and positive energy [27]. This
is called the CPT theorem and up to this day this symmetry seems to hold up experimentally [28].
Another example is the occurrence of symmetry breaking through non-renormalizable eects. Ef-
fective eld theories appear as low-energy approximations to a more overreaching theory, as they only
accurately describe the particles that fall within the energy range considered. Although the eects
of heavier particles do not appear on the low energy regimes, when moving on to higher energies,
symmetries which were assumed on the low end of the energy scale could be broken on the higher
energy theories.
As for the case of spontaneous breaking of symmetries, things are a bit dierent. Instead of the
asymmetry existing in the equations of motion themselves, it occurs instead in one of the possible
solutions to those equations, arising dynamically from the system.
A simple example can be given in order to illustrate this. Consider a cylindrical rod held hori-
zontally. If we let the rod fall, it will spontaneously choose a direction, breaking the initial rotational
symmetry. The state in which the symmetry of the system is broken is one of the innite solutions of
the dynamical equations governing the system.
This same phenomenon can be observed when a ferromagnet is cooled below its critical temperature
(Curie temperature Tc). Initially the system presents no magnetization, as it is has T > Tc, but a net
magnetization emerges as soon as T < Tc, where the spins align spontaneously in a given direction,
breaking the initial symmetry of the system.
In the context of particle physics, the occurrence of spontaneous symmetry breaking is fundamental
in explaining certain phenomena. One example of the breaking of a discrete symmetry is the Yukawa
interaction, which can be explained via the Lagrangian density
L =1
2(∂φ)2 − µ2φ2 − λφ4. (3.3)
where φ is a real scalar eld. This Lagrangian density represents a system consisting of a self-interacting
scalar eld φ, with the potential of the system consisting of the V (φ) = −µ2φ2 − λφ4 terms. It is
invariant under the global Z symmetry φ → −φ or, if we instead adopt a complex scalar eld, the
global U(I) symmetry φ→ φeiα, with α a constant.
If our potential has a minimum value at some point φ0 6= 0, then its symmetry would be eectively
18
3.2 Explicit and Spontaneous Breaking of Symmetries
broken. In this case, by taking µ2 = 0 the potential would have two possible minimum values φ0 =
±√
1/(2λ)µ, which when selected by the system would spontaneously break the underlined global
symmetry.
In the case of continuous SSB, the discovery of the Goldstone's Theorem [29] is fundamental in
explaining several of results in particle physics. As a result of the continuous symmetry being broken,
massless bosons appear (Goldstone bosons), with the number of bosons being equal to the number of
generators of the broken symmetry. These massless bosons are crucial in the understanding the Higgs
mechanism. In this case, the picture changes as the symmetry being broken is local, instead of the
global one described by the last example.
A similar model to the one given by equation 3.3 can be given in order to explain this mechanism.
Consider the Lagrangian of the complex scalar eld φ, φ∗ coupled with the electromagnetic eld
L = −(Dµφ)∗(Dµφ) + µ2φ∗φ− 1
4λ(φ∗φ)2 − 1
4FµνFµν (3.4)
where Dµφ = ∂φ − iqAµφ is the covariant derivative, and the potential can be identied similarly as
V (φ, φ∗) = µ2φ∗φ − 14λ(φ∗φ)2. This potential is what is called the "Mexican Hat Potential", due to
its shape, which can be seen on Figure 3.2.
Figure 3.2: The famous Mexican Hat potential which describes how a system spontaneously breaks its initialsymmetry by rolling onto the minimum of the potential, where a circle of innite possible solutions exist.
The minima is now given by a point in a circle on the complex plane, i.e. |φ|2 = µ2a2/2 with
a =√
4µ2/λ, and so the "rolling" of the eld to one of these values would eectively break the
symmetry.
If we dene the scalar eld φ as,
φ(x) =1√2
(α+ β(x) + iγ(x)) (3.5)
then our kinetic terms are changed into
− (Dµφ)∗(Dµφ) = −1
2(∂β)2 − 1
2(∂γ)2 − 1
2q2α2A2 + qαAµ∂
µβ + ... (3.6)
19
3. Symmetry Breaking
where we end up having a quadratic coupling between the gauge eld A and the scalar γ. With an
appropriate gauge change, given by Aµ(x) = Vµ(x) + (qα)−1∂µγ(x), this can be corrected, as now the
kinetic terms are give by
− (Dµφ)∗(Dµφ) = −1
2(∂β)2 − 1
2q2α2V 2 + .... (3.7)
This is the Higgs mechanism: The gauge change brought with it a new mass term with m = qα,
with the added advantage of absorbing the scalar eld γ and collapsing the potential onto a non-zero
minimum, thus breaking its symmetry.
The goal of this work is not to explain the Higgs mechanism in detail, but rather bring some
light into the relevance that the breaking of spontaneous symmetries has in explain certain physical
phenomena. The gravitational models that will be presented will have some aspects of these SSB in
some sense, which though not representing the same mechanism, are easier to understand having a
more complete conceptual baggage.
3.3 Observer and Particle LSB
Due to the extensive application of this symmetry in physical theories, discussing it without clarify-
ing which exact symmetry is broken brings with it a lot of confusion, as not all Lorentz Transformations
(LT's) are the same. We can group the dierent types into two groups: as Observer and Particle
Lorentz transformations. Following this labeling, we can characterize them in the following manner.
• Observer Lorentz transformations: These are the transformations one usually thinks of
when working in Special Relativity. They relate the observations of two inertial observers which
can have dierent velocities or can be rotated in some manner in relation to each other. Not
having this kind of invariance would mean that the choice of reference frame would alter the
results obtained when measuring particular phenomena, i.e. a particle's mass would vary from
one point to another. Due to this, theories must preserve this symmetry, something which is
accomplished by writing the laws of physics in terms of covariant equations. If one takes an
action S dened on a manifoldM with a metric g and which depends on some eld Φ(x), then
S[g,Φ(x)] =
∫
ML(g,Φ(x),∇Φ(x)) (3.8)
must be a Lorentz scalar in order to be invariant. The equations of motion derived from this
action will be invariant to boosts and rotations between reference frames of the form
xµ → x′µ = Λµνxν (3.9)
with the elds also being transformed accordingly.
• Particle Lorentz transformations: If in the rst case the dierence between dierent exper-
iments was only a change of coordinates, in this case two identical experiments can be boosted
or rotated relative to each other by the same observer. This happens because as we rotate our
20
3.4 Standard Model Extension and LSB
experiment in relation to a certain eld, the relations between them are also altered. In order to
clarify this fact, a couple of examples can be given.
Consider a system composed of a charge of massm and charge q being aected by a perpendicular
magnetic eld ~B. Its motion can be described by
md~v
dt= q~v × ~B, (3.10)
an equation that we know is valid in all reference frames. The particle in this situation would
move in a circular motion in a plane perpendicular to the eld. Suppose we made a particle LT
by means of a boost; the consequence of this would be a larger radius in the trajectory of the
particle, as its momentum would be increased. On the other hand, if we made an observer boost
along the particle's trajectory, the result would be dierent, as a drift in the particle would arise
due to now there being an electric eld.
The other example is also straightforward. Considering a mass on an inclined plane on the surface
of the earth [30], if one simply changes the system of coordinates that describe the system, the
acceleration that the mass obtains is the same in both reference frames, diering only from a
change of coordinates. However, if we rotate the inclined plane, that acceleration would change
perceptively as the direction of the gravitational eld in relation to the xed background would
be dierent (it would be as if we created a new ramp with a more drastic incline).
In conclusion, the type of Lorentz symmetry that is eectively broken in the following sections will
be of the second type, i.e. particle Lorentz transformations.
3.4 Standard Model Extension and LSB
The Standard Model gives us an accurate description of the myriad of phenomena which occur
between the basic particles and forces at very small scales. On larger scales, this burden falls on the
classical description provided to us by General Relativity which, as far as experimental conrmation
goes, has proven itself within the class of physical eects it tries to eectively describe [10].
As these two eld theories are expected to merge at the Planck level, with energies around mP =
1019GeV, into a single unied and consistent description of nature, search for possible signals at this
scale is paramount to achieve a deeper understanding how this merger might theoretically occur. One
possible candidate for these signals is the breaking of Lorentz symmetry.
This work will be centered on a particular type of alternative gravitational theories which are
based on the introduction of vector elds in the Lagrangian density of our system, the so called Aether
theories.
The main motivation for this type of approach emerged from the work done initially by Alan
Kostelecký circa 1988, which consisted in the study of natural Lorentz symmetry breaking (LSB)
mechanisms in bosonic string theory, with the goal of trying to explain the relationship between the
26-dimensional spacetime needed for that theory, and the four at dimensions that we know of, with
21
3. Symmetry Breaking
the assumption that the breaking the symmetry could bring with it the compactication of extra
dimensions [31].
The raise in interest regarding the breaking of fundamental symmetries (such as Lorentz and
CPT violation) in eld theories then culminated on the creation of the Standard-Model extension. A
framework built from the core elements of the Standard Model and General Relativity, as Kostelecký
refers to in Ref. [10]
(...) suppressed eects emerging from the underlying unied quantum gravity theory might
be observable in sensitive experiments performed at our presently attainable low-energy
scales. (...) Any observable signals of Lorentz violation can be described using eective
eld theory. To ensure that known physics is reproduced, a realistic theory of this type must
contain both general relativity and the SM, perhaps together with suppressed higher-order
terms in the gravitational and SM sectors.
To give an idea, the Standard-Model Extension can be represented by an action consisting of a
partial sum of terms given by,
SSME = SSM + SLV + Sgravity + ... (3.11)
where SSM is the SM action (although with some gravitational couplings [10]) with a corresponding
Lagrangian density given by
LSM = Llepton + Lquark + LY ukawa + LHiggs + Lgauge (3.12)
SLV corresponds to the SM Lorentz and CPT-violating terms and Sgravity is the gravity sector of
the Lagrangian.
The Lorentz-violating terms on the SME take the form of Lorentz-violating operators coupled to
coecients which will be dened through Lorentz indexes. The existence of non-zero LV coecients
could appear through various mechanisms, one of which (and the one which will be crucial for the
models referred to herein) is spontaneous Lorentz violation (SLV).
The classication of the Lorentz-violating terms can be done through the observed properties under
CPT [10], as the breaking of this symmetry in Minkowski-spacetime implies Lorentz violation.
The pure gravity action can be written as,
Sgravity =1
2k
∫d4xLgravity (3.13)
with the Lagrangian density consisting of a Lorentz invariant and a Lorentz violating part Lgravity =
LLI + LLV + ... the latter being considered in the limit in which the torsion vanishes [10].
Although it is in this Lorentz violating Lagrangian density that we shall focus our attention, a
brief example follows of how the insertion of Lorentz breaking terms into our Lagrangian density can
change some known physical properties of a system.
Consider the following QED Lagrangian density with isotropic LV [17] which allows us to study LV
in the case of the photons and electron/positrons:
22
3.4 Standard Model Extension and LSB
L = ψ(iγµDµ −m)ψ − 1
4FµνF
µν + ia1ψγiDiψ +
ia2
M2Djψγ
iDiDjψ +a3
4M2Fkj∂
2i F
kj . (3.14)
The Lorentz violating terms are coupled to parameters which allow to test the constraints on some
observable quantities. The rst two terms belong to the usual QED Lagrangian (the Dirac term
along with the electromagnetic energy). The remainder terms are the Lorentz violating ones, which
are coupled to the scale constant M and the dimensionless parameters a1, a2, a3. The u, j indexes
represent the spatial components of the respective terms. From this Lagrangian one can then derive
the dispersion relations, which will naturally depend on the LV parameters, as follows,
E2γ = k2 +
a3k4
M2(3.15)
E2e = m2 + p2(1 + 2a1) +
2a2p4
M2. (3.16)
which are clearly a deviation from the relativistic dispersion relations. The work done in [17] focused on
obtaining constraints for the ai parameters above by studying the interaction of high-energy photons
with the Earth's atmosphere and magnetic eld and how the detection of photon-induced showers with
energies above 1019eV would constrain those parameters to values of the order of |a1| ≤ 10−25, |a2| =|a3| ≤ 10−7, as referred to in chapter 1. The small values for these parameters are in agreement with
the fact that no compelling evidence exists for Lorentz violation.
In the following section a group of models which where created with the goal of trying to obtain
similar constraints will be presented. Focusing on the gravitational sector of the Standard Model
Extension, the aether models present themselves as phenomenological probes to test the existence of
LSB within astrophysical bodies and cosmologies.
23
3. Symmetry Breaking
24
4Vector Theories
Contents
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Aether Theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3 The Bumblebee Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
25
4. Vector Theories
4.1 Introduction
In the following section, a brief introduction to Einstein-Aether theories will be made. Starting
from the dening Lagrangian density of these theories, a brief historical review of how they came into
existence is made. As these theories are relevant in the context of dark matter, particularly as a result of
the emergence of the MOND theories (Modied Newtonian Dynamics) and later on the TeVeS models
which were used as an attempt to explain the phenomenon, the dark matter and energy problem will
also be described. Within the Aether models, two applications will be presented: one regarding the
change in the gravitational constant and the other regarding ination. Finally, the Bumblebee model
will be introduced. This model will later on be used in the work presented in chapter 5 as a way to
obtain astrophysical constraints on the existence of a vector eld with a non-vanishing expectation
value inside spherical star-like bodies.
4.2 Aether Theories
Although some of the more direct references [3235] mention the introduction of Lorentz breaking
symmetries in gravitational theories to the work of Kostelecký [10, 12, 31], the study of gravitationally
coupled vector eld theories dates back to a decade earlier, with the work of Will and Nordtvedt. Their
work, circa 1972, was based on using the Post-Newtonian Formalism (PPN) (see chapter 1 for more
details) to show the eects of a preferred frame of reference in the tides of the earth, the perihelion-shift
of the planets and the variation of the earth's rotation rate [36].
Aether models are based on the introduction of a vector eld in the Lagrangian density of the
system with a non-vanishing vacuum expectation value. Due to that property, the vector eld will
dynamically select a preferred frame at each point in space-time, spontaneously breaking Lorentz
invariance. This is a mechanism reminiscent of the breaking of local gauge symmetry in the Higgs
mechanism as explained in chapter 3 and in [37] and it serves as phenomenological representation of
the LSB terms in the gravitation sector of the SME, as discussed in Ref. [10].
Aether theories have been used, in parallel with the work of Kostelecký [10], as phenomenological
probes of LSB in quantum gravity and as models for ination and dark energy [32] and the corre-
sponding action consists in a 4-vector Aµ coupled to gravity, which can be written as,
S =
∫d4x√−g
[R
16πGN+ LAE(gµν , Aα)
]+ SM (gµν ,Φ) (4.1)
where SM stands for the matter action, gµν the metric and Φ the matter elds which couple only to
the metric [33, 35] and not to Aα.
For the Einstein-aether Lagrangean LAE we can start by writing it in the most general way given
by Ref. [33] (and the references found within),
LAE = Kαβµν − V (AµAµ) (4.2)
with V (AµAµ) being a general potential which depends on our vector Aµ.
The coecients Kαβµν can be given by
Kαβµν = K(1)αβµν +K(2)αβµν +K(3)αβµν (4.3)
26
4.2 Aether Theories
with
K(1)αβµν = (c1gαβgµν + c2δ
αµδ
βν + c3δ
αν δ
βµ + c4A
αAβgµν)(∇αAµ∇βAν) (4.4)
K(2)αβµν = (c5δανA
βAµ + c6gαβAµAν + c7δ
αµA
βAν + c8AαAβAµAν)(∇αAµ∇βAν) (4.5)
K(3)αβµν = c91
2FµνF
µν + c10δαµδ
βνRαβA
µAν + c11RδαµAαA
β (4.6)
in which the c4 and c8 gauge the relevance of the covariant derivatives along Aµ and Fµν = ∂µAν−∂νAµ.
The K(3) is the most relevant of the three, as it contains terms that represent the coupling be-
tween the vector eld and the geometry of the system given by the Riemann tensor (particularly the
c10δαµδ
βνRαβA
µAν and c11RδαµAαA
β terms) and the K(1) and K(2) are simplied to a single kinetic
term β(∇µAµ)2.
As stated previously these models are a subclass of Lagrangian densities where torsion is not
considered, in contrast to the more general case considered by Kostelecký in Ref. [10] and following
work starts with a Lagrangian density consisting of a terms mainly pertaining to the K(3) tensor, as
can be seen in Ref. [35]:
S =
∫d4x√−g
[R
16πG− β1
2FµνF
µν − β(∇µAµ)2 + c10RµνAµAν + c11RAµA
µ − V (AµAµ)
](4.7)
These models did not evolve from a vaccum, but rather from the complete opposite: the existence of
dark matter.
Hints for the existence of dark matter emerged in the early 1930s with the work of J. Oort and
F. Zwicky. Oort, by studying the Doppler shifts of stars in the Milky Way and thus obtaining their
velocities, observed that those stars had enough velocity to escape the gravitational pull provided by
the luminous mass of the galaxy i.e. the mass of the bodies that were directly visible [38].
The easiest way of studying stellar bodies is by measuring how much light they emit per unit of
time, i.e. their luminosity. Because we can accurately measure the mass of the Sun, we calculate its
mass-to-luminosity ratio ML and use it as a standard, which allows us to obtain estimates of the masses
of other astronomical bodies by comparison.
Zwicky used the M/L ratios of the nebulae in the Coma cluster in order to obtain their mass,
and discovered that this was only 2% of the average mass of the nebulae. This value was obtained by
dividing the total mass of the cluster by the number of observed nebulae (about 1000). [39]
Another way in which this mass discrepancy was observed was by assuming that galaxies in galaxy
clusters behaved like planets in the Solar System, with a velocity dispersion given by the Newtonian
expression
v(r) =
√GNm(r)
r. (4.8)
where m(r) is the mass enclosed within a sphere of radius r. Vera Rubin et al. [40] studied the rotation
curves of 60 galaxies and compared the results with what would be obtained by the expression above.
She found that the rotation curves of those galaxies where "at", meaning that for larger radii the
velocities of the galaxies increase until a threshold is reached and remain relatively constant for larger
27
4. Vector Theories
radii. This is unexpected because if one considers that only the luminous mass in the center of the
cluster as the only source of gravitational pull m(r) = M , the velocity should be higher closer to that
center and lower in the periphery, where the gravitational pull is smaller (as can be immediately seen
by the expression above). This then implied that that there is some sort of extended non-luminous
matter in the periphery of the galaxy that exerts a gravity pull large enough to keep the galaxies
rotating at those speeds.
Pieces of evidence from other sources also exist: X-ray radiation from the hot gas surrounding
galaxies was used by Vikhlinin et al. [41] to determine the mass distribution of the galaxies; Ratios of
baryonic to total mass in the order of 13% where found, again indicating the existence of extra mass
in those galaxies.
Gravitational lensing also provides some conrmation for these results, and some consider it as
a direct observation of dark matter, as it does not depend on the dynamics of the clusters [42].
Gravitational lensing consists in measuring masses through the deection of light presents when passing
through a gravitational eld, creating one of more images of the original object in a dierent location
where the object would be directly observed. Weak gravitational lensing (where a single image is
formed) has been used to infer the existence of dark matter in the Bullet Cluster [42], where the
matter distribution was found to be dislocated from the cluster's luminous center of mass. This
discrepancy can be seen in Figure 4.1, where the two sets of contour lines (blue and magenta as seen
in color) which represent the observed gravitational mass are located away from the larger physical
bodies (in black) that are part of the cluster.
Figure 4.1: F606W-band image of the Bullet Cluster. Image taken from [43]
As a nal example, Cosmic Microwave Background (CMB) radiation can also be used as a test for
inferring the existence of dark matter, through the measurement of the density parameters present in
the Friedman-Robertson-Walker model. Starting with the Friedman equation,
R2 − 8πG
3ρR2 = −kc2 (4.9)
28
4.2 Aether Theories
which describes a at, closed or open universe for k = 0,+1,−1, we can rewrite it by considering the
densitiy parameter given by
Ω =ρ
ρc=
8πGρ
3H2(4.10)
with H the Hubble parameter. This allows us to obtain the relation
H2(Ω− 1) =kc2
R2, (4.11)
which directly shows a relation between the matter density and the resulting curvature. The limits Ω =
1,Ω > 1,Ω < 1 make the correspondence between a at, closed and open Universe. CMB anisotropy
measurements allow us to obtain values for the total density parameter Ωtot = Ωm+Ωrad+ΩΛ and for
the isolated Ωm = ΩBaryonic+Ωnon−Baryonic parameter. Various measurements have been made, with
the most recent belonging to the Planck 2013 experiment [44]. Its results where Ωmh2 = 0.1423±0.0029,
Ωbaryonic = 0.02207 ± 0.00033 and for non-baryonic cold dark matter Ωch2 = 0.1196 ± 0.0031 (with
h = 6.2606957(29) × 10−34). These results are in agreement with previous experiments and show a
great disparity in the relative percentages of baryonic to non-baryonic matter, again indicating a strong
presence of dark matter in our Universe.
Numerous candidates have been proposed to ll the role of dark matter. Standard model neutrinos
along with sterile neutrinos [45] (same as Standard Model neutrinos but without weak interactions)
have been considered as candidates due to their weak interaction with baryonic matter. One of the
main problems in this assumption is that their abundance in the Universe does not allow them to be
the dominant component of dark matter. [45]
Axions, a result of CP violation physics postulated by Peccei and Quinn [46], have also been
considered, with searches currently being done by the Axion Dark Matter Experiment [47].
Another class of candidates encompasses the super-symmetric candidates emerging from SUSY.
Neutralinos have been widely studied as cold dark matter candidates due to being heavy stable particles
with coupling strengths in the order of the weak interaction [45]. Experiments such as the Cryogenic
Dark Matter Search [48] and more recently the Large Underground Xenon (LUX) [49] seek to detect
these kinds of weakly interacting massive particles (WIMPs) but the results don't seem to strengthen
the dark matter hypothesis.
For more comprehensive reviews of particle candidates to dark matter please check [45, 50, 51].
Besides particle candidates, other alternatives exist to the dark matter hypothesis. Modied New-
tonian Dynamics is one of such alternatives and it consists in a modication of Newton's laws in order
to explain the at rotation curves discussed earlier.
The main assumption in MOND theories is that Newton's law is modied with a dependence in the
acceleration of the system. For low accelerations we have, instead of the usual second law of motion:
~F = m~aµ(x) (4.12)
where µ(x) is any function exhibiting the asymptotic behaviors
µ (x) =
x if x 1
1 if x 1.(4.13)
29
4. Vector Theories
For small values of the ratio aa0
we can compare the acceleration to what we would normally have,
i.e.,
~F = m~g = m~aa
a0→ a =
√ga0. (4.14)
So, for the simple case of a body in an orbital motion around a central mass M, where g = GMr2 we
would have a centripetal acceleration given by a = v2
r so our velocity would be
v4 = GMa0, (4.15)
which would be dependent from the central mass M. The parameter a0 can now be tted to the results
obtained in the velocity curves yielding a0 = 1.2×10−10m/s2 [52]. One of the most interesting aspects
of this theory is that the value of a0 obtained from observations is in the same order of magnitude of
the Hubble constant, a0 ∼ cH0 [53].
Although this framework allows us to explain some of the velocity curves of individual galaxies,
some problems do exist. As Sanders and McGaugh explain [53],
There have also been several contributions attempting to formulate MOND either as a
covariant theory in the spirit of General Relativity, or as a modied particle action (modied
inertia). Whereas none of these attempts has, so far, led to anything like a satisfactory
or complete theory, they provide some insight into the required properties of generalized
theories of gravity and inertia.
One example of an attempted merger of both MOND and General Relativity is what is called the
Tensorial-Vector-Scalar theory, proposed by Bekenstein in 2004 [54]. The assumptions of the theory
are also its main source of criticism [55], as it assumes the existence, besides the metric gαβ and the
matter elds φi, one extra vector eld Uα and two extra scalar elds σ and φ, the rst of which lacks
of a convincing justication due to not having dynamic terms in the action. The action of this theory
also goes in the inverse direction of one of the main aspects of MOND that made it so appealing,
as its simplicity is completely lost, as one can see by the action S = SG + SV + SS + SM with the
corresponding terms given by
SG =
∫R(−g)1/2d4x (4.16)
SV = − K
32πG
∫ [gαβgµνU[α,µ]U[β,ν] −
2λ
K(gµνUµUν + 1)
](−g)1/2d4x (4.17)
SS =1
2
∫ [σ2hαβφ,αφ,β +
G
2l2σ4F (kGσ2)
](−g)1/2d4x (4.18)
SM =
∫L(φi)(−g)1/2d4x. (4.19)
Another point in which MOND theories fall short is the inability to solve some discrepancies in
particular matter dispositions of galaxy clusters, as discussed in [5658]. The matter distributions of
visible matter do not correlate exactly with the non-visible part, an eect which can be seen directly
in the Bullet Cluster, where two clusters of galaxies collided to form a matter distribution with two
distinct peaks of visible matter density, and two other zones where the gravitational lensing eect is
stronger, these being located both far away from each other and far relatively far away from the visible
density peaks.
30
4.2 Aether Theories
4.2.1 Eects on the Gravitational Constant
In this section the goal is to give an overview of some interesting aspects and results obtained by
applying these types of models to the study of cosmology. Starting by selecting the relevant terms for
our Lagrangian, we have
L = K(1)µναβ − λ(uµuµ +m2). (4.20)
where all but the c4 coecients are non-zero and our potential is a Lagrange multiplier eld associ-
ated with the variable λ, for which we will later on deduce its equations of motion. Following the
methodology presented by Ref. [59] and [60], we dene a new tensor given by,
Jµα = Kµναβ∇νuβ (4.21)
which allow us to obtain the equation of motion from the action above, with respect to the vector eld
uµ
∇µJµν = λuν . (4.22)
The constraint condition for this equation is
uµuµ = −m2 (4.23)
which basically says that we are solving our equations of motion in the vacuum expectation value of
our eld.
Given the adopted metric signature (−,+,+,+), we require that the vector uµ is timelike, which
implies that m2 > 0.
Multiplying the λ equation by uν on both sides gives
λ = − 1
m2uν∇µJµν . (4.24)
The metric that we will be using is the FLRW metric, which describes a spatially and isotropic
universe.
ds2 = −dt2 + a2(t)
(1
1− kr2dr2 + r2dΩ2
)(4.25)
The condition of spatial isotropy implies that our Aether vector eld must have only a temporal
component. Because of the condition gνµuµuν = −m2, the vector must be something like
uµ = (m, 0, 0, 0). (4.26)
Inserting this in equation 4.22 we get,
λ(t) = −3(c1 + c2 + c3)H2 + 3c2a
a(4.27)
which depends directly on the aether coecients dened in K(1) term of the previous section.
Besides this, another useful expression is that of the stress energy tensor which can be obtained
by varying the action with respect to the metric gµν . The calculations, albeit cumbersome, are spread
out in the literature related to these models [61], and its expression can be simplied to [59]:
31
4. Vector Theories
Tµν = 2c1(∇uβ∇νuβ −∇βuµ∇βuν)− [∇β(u(µJβν))+
∇β(uβJ(µν))−∇β(u(µJβν))]−
2
m2uα∇βJβαuµuν + gµνL.
(4.28)
The next step is to analyze how the introduction of the aether vector eld aects gravity. The rst
assumption is to consider that we have a system composed of both matter and the vector eld, which
can be expressed by the Einstein eld equations by the sum of the two stress-energy tensors
Rµν −1
2Rgµν = 8πGN (Tmatterµν + T aetherµν ). (4.29)
The gravitational constant GN is the one dened in the Aether Lagrangian from equation 4.1.
Matter is assumed to behave as a perfect uid, with a matter-energy tensor:
Tmatterµν = (ρm + pm)ηµην + pmgµν (4.30)
The Aether eld can be described in the same fashion, dening the corresponding pressure and
density
ρae = −3αH2 , pae = α
[H2 + 2
a
a
], (4.31)
with H = aa the Hubble parameter and α = (c1 + 3c2 + c3)m2 [59].
The Einstein equations lead, through the Bianchi identities, to the energy conservation equation
for a general density ρ and pressure p
ρ+ 3H(ρ+ p) = 0 (4.32)
If we use the denitions above for the aether density and pressure and insert it in equation (31),
we have,
˙ρae + 3H(ρae + pae) =
−3α2HdH
dt+ 3H(−3αH2 + α+ 2α
a
a) =
−3α2H(a
a−H2) + 3H(−2αH2 + 2α
a
a) =
−6αH(a
a−H2) + 6αH(
a
a−H2) = 0,
(4.33)
which proves that the selected terms are a good choice.
For the FLRW metric we have
R00 −1
2Rg00 =
3
8πGN
(H2 +
k
a2
)= T00 = ρae + ρm. (4.34)
The spatial components are given by
Rik −1
2Rgik =
1
8πGN
(H2 + 2
a
a+
k
a2
)= Tik = pae + pm (4.35)
Rewriting these equations in a form similar to the Friedmann equations, substituting the Aether
terms for pressure and density and assuming a at space k = 0, we get
H2 =8πGc
3ρm, (4.36)
a
a= −4πGc
3(ρm + 3pm) (4.37)
32
4.2 Aether Theories
with the denition of a new gravitational constant given by
Gc =GN
1 + 8πGNα. (4.38)
The aether energy density suggests that for a positive energy density, we must have α < 0; positivity
of H2 further implies that 1/(8πGN ) ∼M2p < α < 0 (where Mp ∼ 1019GeV is the Planck mass). This
would imply that our gravitational constant would increase in relation to the original G∗. Because the
acceleration equation (36) depends linearly on Gc, this implies that the net eect would be that the
acceleration rate of the expansion of the universe would become larger.
4.2.2 Ination
Based of the Lagrangian density used in the study above, one can also study the role that these
models play on ination. For this case the potential chosen by Kanno & Soda [62] has a Lagrange
multiplier potential similar to the one used above, where m2 = 1. The ci terms adopted are also
dierent, along with the expression for the timelike vector eld uµ
uµ = (1, 0, 0, 0) . (4.39)
The inital action is of a Scalar-Vector-Tensor theory, where besides the vector eld, we have a
scalar eld φ coupled to it. This new Lagrangian can be understood as a perturbation of the scalar
eld ination model, where a Lorentz violating vector eld is introduced. The inital action is given by
S =
∫d4x√−g
[1
16πGR− β1(φ)∇µuν∇µuν − β2(φ)∇µuν∇νuµ − β3(φ)(∇µuµ)2
−β4(φ)uµuν∇µuα∇νuα + λ(uµuµ + 1)− 1
2(∇φ)2 − V (φ)
] (4.40)
and one can immediately see that the terms pertaining solely to the scalar eld, are those used in the
most simple models that describe ination.
The metric used is
ds2 = −dt2 + a2(t)δijdxidxj . (4.41)
where a = eα.
Inserting the vector eld and the metric in the action gives[62],
S =
∫dt
1
Ne3α
[− 3
8πG
(1 + 8πGβ
)H2 +
1
2φ2 − V (φ)
]. (4.42)
where β = β1 + 3β2 + β3.
From this action the procedure is similar to what was done in the previous section, as we obtain
the relevant equations of motion for the variables of our system β, φ,H. They are, respectively
γH2 =1
3
[H2φ′
2β+V
β
](4.43)
γH ′
H+φ′2
2β+β′
β= 0 (4.44)
φ′′ +H ′
H+ 3φ′ +
V,φH2
+ 3β,φ = 0 (4.45)
33
4. Vector Theories
where Q = dQdη
dηdt with Q′ = dQ
dη and γ = 1 + 18πgβ .
The existence of Lorentz violation implies that the terms relating to it in our model are compar-
atively big to the scalar eld ones. This assumption, i.e. β >> 1&β >> f(φ, φ′′, β′) will imply that
γ → 1 and the rst and third equations will now be
H2 =V
3β(4.46)
φ′ +V,φ3H2
+ β,φ = 0 (4.47)
The choice of potential is that of a parabola, in order to nd the similarities between these conditions
and the slow roll conditions one forces when studying independent scalar eld ination. Besides that,
the authors suggest a quadratic coupling of the scalar eld to the Aether parameters β, so we have
β = ηφ2 ; V =1
2m2φ2. (4.48)
These conditions, along with equations (45) and (46) allow us to obtain the solutions
φ(α) =φ0
a4η(4.49)
H2 =m2
6η(4.50)
which represents an inationary model as a consequence of Lorentz violating parameters in our model.
This can be seen directly if we consider the way in which these parameters where choosen.
Because H = α, we dened a(t) = e2α(t) then α =˙a(t)
a(t) which is the denition of the Hubble
parameter. The deceleration parameter q can be related to H by
H
H2= −(1 + q)⇔ q = −
(1− H
H2
). (4.51)
Because H is a constant we have q = 1, which is the necessary condition for our theory to have
ination.
4.3 The Bumblebee Model
A model that contains a vector eld which dynamically break Lorentz symmetry is called a Bumble-
bee model. These models, although with a simpler form, contain interesting features such as rotations,
boosts and CPT violations.
This subclass of aether models posits the following action functional,
SB =
∫d4x
[1
16πG(R+ ξBµBνRµν) −1
4BµνBµν − V (BµBµ ± b2)
], (4.52)
where the Bµ is the same as the vector eld Aµ (but now following the KS convention), and Bµν being
the Bumblebee eld strenght given by
Bµν = ∇µBν −∇νBµ. (4.53)
The parameter ξ, with dimensions of M−2, represents the coupling between the Ricci tensor and
the Bumblebee eld Bµ and V the potential of the Bumblebee eld which, as in the case of the aether
34
4.3 The Bumblebee Model
theories, is the term that drives the breaking of the Lorentz symmetry of our Lagrangian by collapsing
onto a non-zero minimum at BµBµ = ∓b2. Here, Bµ is one of the Lorentz breaking coecients
referred in chapter 3. The presence of this coecient implies a preferred direction is selected at a
certain Lorentz frame, which implies that the equivalence-principle is locally broken for that particular
frame. Observations of Lorentz violation can emerge if particles or elds interact with the Bumblebee
eld [10].
It is worthy to repeat that within this local frame of reference, local particle Lorentz transformations
(referred to in chapter 3) can be performed without changing the local Bumblebee eld, as under these
transformations Bµ behaves as a set of four scalars. Rotations and boosts that change the local Lorentz
frame (observer Lorentz transformations) allow us to choose arbitrarily the local Lorentz frame being
observed, as under these transformations the Bumblebee eld behaves covariantly as a four-vector.
This allows us to maintain local Lorentz invariance despite having local particle Lorentz violation.
As referred in Ref. [34], a smooth quadratic potential of the form,
V = A(BµBµ ± b2)2 (4.54)
with A a dimensionless constant, is chosen. This potential allows both Nambu-Goldstone excitations
(massless bosons) besides the massive excitations for the cases of V = 0 and V 6= 0 respectively [10].
The other case mirrors one of the presented potentials in the aether model section above, i.e. a
linear Lagrange-multiplier potential which takes the form
V = λ(BµBµ ± b2). (4.55)
Both cases where studied from the particle physics point of view and, besides the spontaneous
lorentz breaking, these potentials present also the breaking of the U(1) gauge invariance and other
implications to the behavior of the matter sector, the photon and the graviton. A good review of
experimental proposals to test the result of Bumblebee models can also be found in [63].
Notice that the potential V is assumed to depend on BµBµ ± b2 with b 6= 0 the non-vanishing vev
signalling the spontaneous Lorentz symmetry breaking.
Variation of Eq. (4.52) with respect to the metric yield the modied equations of motion [10],
Rµν −1
2Rgµν = 8πG(TMµν + TBµν), (4.56)
where TMµν is the matter stress-energy tensor and TBµν the Bumblebee stress-energy tensor, dened
as
TBµν ≡ −BµαBαν −1
4BαβB
αβgµν − V gµν + 2V ′BµBν +ξ
8πG
[1
2BαBβRαβgµν −BµBαRαν
+1
2∇α∇µ(BαBν) +
1
2∇α∇ν(BαBµ)− 1
2∇2(BµBν)− 1
2gµνDα∇β(BαBβ)
].
(4.57)
The equations for the Bumblebee eld are
∇µBµν = 2V ′Bν − ξ
8πGBµR
µν , (4.58)
where the prime represents derivative in respect to the argument.
35
4. Vector Theories
No separate conservation laws are assumed for matter and the Bumblebee vector eld in the work
that will be presented in the following chapter. The covariant (non)conservation law (which is not
used) can be obtained directly from the Bianchi identities ∇µGµν applied to both sides of the modied
eld equations (5.7): this leads to ∇µTMµν = −∇µTBµν 6= 0, which may be interpreted as an energy
transfer between the Bumblebee and matter.
Studies using these models have recently emerged in the literature. The vacuum solutions for the
Bumblebee eld for purely radial, temporal/radial and temporal/axial Lorentz symmetry breaking
where obtained in [64]. For the rst case, a new black-hole solution was found where its Schwarzschild
radius presents itself with a slight perturbation. The second case was analyzed through the PPN
formalism where a set of PPN parameters was obtained. The nal case, due to a breaking of isotropy,
was not possible to analyze directly through the PPN formalism, although an estimation of the PPN
parameter γ was obtained.
Other work which directly deals with these models was done by Bluhm in Ref. [34], where the
possibility of a Higgs mechanism was analyzed. Studies referring to the electrodynamics of these elds
was done in Ref. [65], where the Bumblebee eld was interpreted as a photon eld and its propagation
velocity was studied, along with its implications on acelerator physics and cosmic ray observations.
The following chapter will feature original work which closely follows the work done in Ref. [1].
In it, these Bumblebee models are used in order to constrain the possibility of Lorentz violating elds
existing in astrophysical bodies such as the Sun.
36
5Astrophysical constraints on the
Bumblebee
Contents
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2 Static, spherically symmetric scenario . . . . . . . . . . . . . . . . . . . . . 38
5.3 Perturbative Eect of the Bumblebee Field . . . . . . . . . . . . . . . . . 39
5.4 Numerical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
37
5. Astrophysical constraints on the Bumblebee
5.1 Introduction
This chapter presents the results of a perturbation induced by the Bumblebee eld on a system
like the Sun. It follows the original work presented in Ref. [1].
5.2 Static, spherically symmetric scenario
Given that the relevant quantities such as the density, pressure and scalar curvature inside a
spherical symmetric body such as the Sun present a very strong radial variation in comparison with
the temporal component, the Bumblebee eld is chosen to be
Bµ = (0, B(r), 0, 0). (5.1)
Accordingly, a static Birkho metric is also selected,
gµν = diag
[−e2ν(r),
(1− 2Gm(r)
r
)−1
, r2, r2 sin2 θ
]. (5.2)
where m(r) is the mass prole in function of the radial coordinate, and it is assumed that the potential
takes a quadratic form, for simplicity,
V = A(BµBµ − b2)2, (5.3)
with the adopted sign reecting the spacelike nature of the Bumblebee eld.
For the radial case µ = r, the Ricci tensor is given by,
Rrr =G(m′r −m)(2 + rν′)
r2(r − 2Gm)− (ν′)2 − ν′′. (5.4)
The only non-vanishing component of Eq. (4.58) is for the component µ = r, 16πGV ′grr − ξRrr = 0
which yields,
2A(grrB2 − b2) =
ξ
16πGr3
[Gr2[m′ν′ + 2m(ν′′ + ν′2)] +Gr(2m′ −mν′)− 2mG− r3(ν′′ + ν′)
]. (5.5)
Through some algebraic manipulation we can the calculate the bumblebee eld strength,
B2 =
(1− 2Gm
r
)[b2 +
ξ
32πGAr3(Gr2[m′ν′ + 2m(ν′′ + ν′2)] +Gr(2m′ −mν′)− 2mG− r3[ν′′ + ν′])
].
(5.6)
In order to obtain the pressure and density equations, the trace-reversed eld equations are con-
sidered, giving,
Eµν ≡ Rµν − 8πG
[TMµν + TBµν −
1
2gµν(TM + TB)
]= 0, (5.7)
with TM and TB the traces of the stress energy tensors for normal matter and the Bumblebee eld,
respectively.
The stress-energy tensor for normal matter in the perfect uid hypothesis is given by,
TMµν = (ρ+ p)uµuν + pgµν , (5.8)
38
5.3 Perturbative Eect of the Bumblebee Field
where uµ is the four velocity; in the static scenario and given that uµuµ = −1, we have uµ = (e2ν(r),~0),
so that
TMµν = diag
(e2νρ,
p
1− 2Gmr
, r2p, r2p sin2(θ)
), (5.9)
with trace T = 3p− ρ.If we then consider the following combinations of the trace-reversed eld equations,
gttEtt − grrErr = 0,
gθθEθθ = 0.(5.10)
we can derive the equations that will allow us to obtain p(r), ρ(r) and ν(r). Without the Bumblebee
eld, these quantities (denoted with the subscript 0) are simply
p0(r) =r(r − 2Gm0)ν′0 −Gm0
4πGr3, (5.11)
ρ0(r) =m′0
4πr2, (5.12)
ν′0(r) = Gm0 + 4πp0r
3
r(r − 2Gm0), (5.13)
which, along with a state equation that relates p0 and ρ0, yields a closed set of four dierential equations
with four unknowns.
In the presence of the Bumblebee eld, the related eld equation (4.58) also has to be included;
solving Eq. (5.7), the pressure and density are then given by
p(r) =1
8πGr4
[rξB(r − 2Gm)2B′(2 + rν′) + r(8πGV r3 − 2Gm+ 2r(r − 2Gm)ν′)−
B2(r − 2Gm)(−2ξGm(−1 + r(ν′(−2 + rν′) + rν′′)) + r[16πGV ′r2
−ξ + rξ(ν′(−2 + rν′) + rν′′)])] ,
(5.14)
and
ρ(r) =1
8πGr4[−r2(8πGV r2 + ξ(r − 2Gm)2B′2 − 2Gm′) + rξB(r − 2Gm)(B′(−4r + 3Gm
+5Grm′)− r(r − 2Gm)B′′) + ξB2(3G2m2
−2G2rm(3m′ + rm′′) + r2[−1 +G(m′(4−Gm′) + rm′′)])].
. (5.15)
Although we have a complete set of equations that describe the behaviour of our system, the
solution of that set of equations implies very intensive numerical computations. This is expected as
we have up to second order derivatives on the ν(r) function inside the expression for the Bumblebee
eld, which itself appears inside both the pressure and density equations as a second derivative. The
combined expansion of all the terms in these expressions, combined with the fact that we would have
a fourth order system in our hands, brings a great deal of diculty to the problem. The linearisation
of the system of equations (not shown) does not simplify the problem signicantly, as the number of
terms in the equation would increase immensely.
5.3 Perturbative Eect of the Bumblebee Field
Since the stellar structure of the Sun is known to be well described by General Relativity, we
consider the perturbation to be of zero order, i.e. we replace the quantities on the r.h.s. of the
39
5. Astrophysical constraints on the Bumblebee
equations mentioned above by the unperturbed expressions for m0(r) and ν0(r) and the Bumblebee
eld. Regarding the latter, it is more straightforward to resort instead to Eq. (5.5), since at zeroth
order one has
Rrr = 8πG
(TMrr −
1
2grrT
M
)= 4πG(ρ0 − p0)grr, (5.16)
which leads to the following expression for the Bumblebee eld equation
B2(r) =
(1− 2Gm
r
)(b2 +
ρ0 − p0
8Agrr
). (5.17)
Since the unperturbed solutions ρ0(χ) and p0(χ) vanish at the boundary of the spherical body, the
above shows that the Bumblebee eld collapses onto its VEV as it crosses to its outer solution (where
TMµν = 0), BµBµ = b2. This is consistent with the approach followed in Ref. [64], where the latter
condition was also assumed.
Following the changes above, the expressions for the pressure and density can now be obtained,
giving
p(r) = p0 + [ξ(p− ρ)]2 +ξ2(ρ′ − p′)
2AGπr
(1− 2Gm
r
)3
(2 + ν′r) +ξ
Aπr3
(1− 2Gm
r
)2
[8Ab2
+ξ(ρ− p)]×[(2 + ν′r)(m−m′r) +
r
G
(1− 2Gm
r
)(1 + r(4Gπpr + 2ν′
−r[4Gπρ+ (ν′)2 + ν′′])− 2Gm
r[1 + r(ν′[2− ν′r]− rν′′)]
)],
(5.18)
ρ(r) = ρ0 −[ξ
8(ρ− p)
]2
+ξ2
128AGπr
(1− 2Gm
r
)2 [(4 +G
[mr− 9m′
])(p′ − ρ′)
+
(1− 2Gm
r
)3
(p′′ − ρ′′)r]
+ξ
64AGπr2
(1− 2Gm
r
)[8Ab2
+ξ(ρ− p)][2
(Gm
r
)2
+ 6Gm′(1−Gm′) + 2Gm′′r − 1− 2Gm
r(1 + 2Gm′′r)
],
(5.19)
The advantage of considering the simplistic model provided by the polytropic EOS Eq. (2.25) lies
in the possibility of rewriting the rather convoluted expressions above only in terms of the LE solution
θ(χ). For this, we now introduce the dimensionless parameters
α ≡ ξ2
R2G, β ≡ ξ3b2
R2G, γ ≡ Rs
R, (5.20)
where Rs ≡ 2GM is the Schwarzschild radius of the star, together with the form factor
φ ≡ 3M
4πρcR3, (5.21)
and the EOS parameter ωc ≡ pc/ρc.Using the relations (2.28) and the expression for ν′0(r) from Eq. (5.11), we can obtain
ν′0(χ) =3γ(χωcθ
1+n − θ′)2φχ2
f + 6γχθ′, (5.22)
and the form factor becomes φ = −3θ′(χf )/χf displaying the homology invariance of the LE Eq.
(2.31).
40
5.4 Numerical analysis
It is now possible to rewrite the above expressions for the pressure and density in terms of the
LE solution θ(χ) and its derivatives. This allows us to obtain a more manageable form: separating
the contributions to the pressure and density arising from the non-vanishing vev b and the potential
strength A as
p(χ) = p0(χ) + pb(χ) + pV (χ) + δ(χ), (5.23)
ρ(χ) = ρ0(χ) + ρb(χ) + ρV (χ)− δ(χ), (5.24)
we have
pb(χ) =β
16πφ3ξ2χ4fχ
(φχ2f + 3γχθ′)2 × [3γχ2θn(−1 + ωcθ)− (5.25)
2(−1 + χ[ν′(−2 + χν′) + χν′′])(φχ2f + 3γχθ′) + 3γχ(2 + χν′)(θ′ + χθ′′)],
ρb(χ) =β
16πφ3ξ2χ4fχ
(φχ2f + 3γχθ′)× [2φ2χ4
f + 3γχ(45γχθ′2 + (5.26)
χ[14φχ2fθ′′ + 9γχ2θ′′2 + 2φχ2
fχθ′′′] + 2θ′[7φχ2
f + 3γχ2(10θ′′ + χθ′′′)])],
pV (χ) =3γα2θn
1024π2Aφ4ξ2χ4fχ
2(φχ2
f + 3γχθ′)2 ×[(ωc − 1)(3γχ2θn[1− ωcθ] + (5.27)
2[−1 + χ(ν′[χν′ − 2] + χν′)](φχ2f + 3γχθ′))− χ
θ(2 + χν′)(θ′[θ([1 + n]φχ2
fωc + 3γ[ωcθ − 1]) +
3γχ([1 + n]ωcθ − n)θ′ − nφχ2f ] + 3γχθ[ωcθ − 1]θ′′)
],
ρV (χ) =3γα2θn
2048π2Aφ4ξ2χ4fχ
2(φχ2
f + 3γχθ′)×[χ
θ2(φχ2
f + 3γχθ′)(2[1− n]nχθ′2[φχ2f + 3γχθ′] + (5.28)
nθ[θ′(χθ′[−51γ + 2(1 + n)φχ2fωc + 6γ(1 + n)χωcθ
′]− 8φχ2f )−
χ(2φχ2f + 33γχθ′)θ′′] + [1 + n]ωcθ
2[θ′(8φχ2f + 51γχθ′) + χ(2φχ2
f + 33γχθ′)θ′′])−
2(2φ2χ4f + 3γχ[45γχθ′2 + χ(14φχ2
fθ′′ + 9γχ2θ′′2 + 2φχ2
fχθ′′′) +
2θ′(7φχ2f + 3γχ2[10θ′′ + χθ′′′])]) + 2ωcθ(2φ
2χ4f + 3γ + χ[45γχθ′2 +
χ(14φχ2fθ′′ + 9γχ2θ′′2 + 2φχ2
fχθ′′′) + 2θ′(7φχ2
f + 3γχ2[10θ′′ + χθ′′′])])
],
and
δ(χ) =9γ2α2θ2n(ωcθ − 1)2
4096π2φ2ξ2. (5.29)
The latter appears in both the pressure and density perturbations and, as shall be shown in the
following section, has a negligible impact when compared with the remaining contributions.
5.4 Numerical analysis
In Fig. 2, the prole of the contributions to Eq. (5.23) was shown for values between 0.1 and 10
solar masses. The radii varied between 0.1 and 1000 times those of the Sun. The motivation behind
having a much bigger variation in the radii in comparison to the masses stems from the fact that this
is what is observed in the set of chosen stars. Stars that have a big variance in radius show relatively
low variance in their mass. This can be seen by the chosen set of known stars in Table 5.1:
41
5. Astrophysical constraints on the Bumblebee
−28
−23
−18
−13
−8
−3
0.1R
R
1000R
logpbp0
ξb2 = 10−27, A = 10−11
n = 2.9
n = 3
n = 3.1
0 1 2 3 4 5 6−28
−23
−18
−13
−8
−3
0.1R
R1000R
χ
log
∣ ∣pVp0
∣ ∣
n = 2.9
n = 3
n = 3.1
−28
−23
−18
−13
−8
−3
0.1R
R
1000R
logρbρ0
ξb2 = 10−22, A = 10−6
n = 2.9
n = 3
n = 3.1
0 1 2 3 4 5 6−28
−23
−18
−13
−8
−3
0.1R
R
1000R
χ
log
∣ ∣ρVρ0
∣ ∣
n = 2.9
n = 3
n = 3.1
−28
−23
−18
−13
−8
−3
0.1MM
10M
logpbp0
ξb2 = 10−27, A = 10−11
n = 2.9
n = 3
n = 3.1
0 1 2 3 4 5 6−28
−23
−18
−13
−8
−3
0.1MM
10M
χ
log
∣ ∣pVp0
∣ ∣
n = 2.9
n = 3
n = 3.1
−28
−23
−18
−13
−8
−3
0.1MM
10M
logρbρ0
ξb2 = 10−22, A = 10−6
n = 2.9
n = 3
n = 3.1
0 1 2 3 4 5 6−28
−23
−18
−13
−8
−3
M
χ
log
∣ ∣ρVρ0
∣ ∣n = 2.9
n = 3
n = 3.1
Figure 5.1: Prole of the relative perturbations pb/p0, pV /p0, ρb/ρ0 and ρV /ρ0 induced by the Bumblebee.The parameters ξb2 and A were chosen so that the maximum of the perturbations reaches the adopted 1%limit, showed by the horizontal line in each plot.
42
5.4 Numerical analysis
Star Name Mass Radius
Wolf 359 0.09M 0.16RBetelgeuse 7.7− 20M 950− 1200RAntares 12.4M 883RVY Canis Majoris 17M 1420R
Table 5.1: Selected stars that used as models for the numerical analysis of the Bumblebee perturbation.
The abrupt variations in the perturbations for χ ≈ 1 and χ ≈ 2.7 are the result of the prole
changing between positive and negative values around those points and the adopted logarithmic scale.
The variation of the polytropic index n with the size of the star is also shown.
The values of the parameters (ξ, b, A) are chosen so that the maximum of the relative perturbations
is of the order of 1%, the order of magnitude of the current accuracy of the central temperature of the
Sun [6668]. For reference, the values of these parameters for ξ = 10−11, b = 10−8 are shown in table
5.2 for all the model stars considered in the numerical analysis.
−20 −10 0 10 20−30
−25
−20
−15
−10
−5
log(ξ[GeV 2])
log(
bGeV)
pbρb
Figure 5.2: Allowed region (in grey) for a relative perturbation of less than 1% for pb and ρb.
A small variation in the polytropic index does not cause signicant changes on the obtained bounds
for the system: in particular, n does not impact signicantly the value of ρb, as can be seen directly in
the equation for ρb. If we increase the radius (thus lowering γ and α) the impact of the non-vanishing
vev increases, while leading to a lower contribution from the potential term. If we increase the mass,
we end up having smaller eects on all quantities except ρV , which is rather insensitive to variations
of M .
If we x M = M and R = R, we can then nd values for the parameters of the model (ξ, b, A)
that lead to relative perturbations of less than 1%. The allowed parameter space can be obtained, as
depicted in Fig. 3. Notice that the allowed values for A are bounded from below, since this quantity
appears in the denominator of pV and ρV ; conversely, the region allowed for ξb2 is bounded from above.
Having said this, we can now obtain the bounds for the parameters of our model,
ξb2 . 10−23 ,ξ√A
GeV2 . 10−3 → ξ√AG
. 1034. (5.30)
43
5. Astrophysical constraints on the Bumblebee
−20 −18 −16 −14 −12 −10 −8 −6−30
−20
−10
0
log(ξ[GeV 2])
log(A
)
pVρV
Figure 5.3: Allowed region (in grey) for a relative perturbation of less than 1% for pV and ρV .
It is also worth noting that for the considered sets of parameters, the term δ(χ) is negligible in
comparison with the other terms in both the pressure and density equations, as mentioned after Eq.
(5.29): indeed, one can calculate numerically that δ . 10−34pV for the considered masses and radii.
M,R α× 10−33 β × 10−49 γ ωc0.1 M, R 1.18 1.18 4.07× 10−7 6.946× 10−7
M, R 1.18 1.18 4.07× 10−6 6.946× 10−6
10 M, R 1.18 1.18 4.07× 10−5 6.946× 10−5
M, 0.1 R 1.18 1.18 4.07× 10−5 6.946× 10−5
M, 1000 R 1.18 1.18 4.07× 10−9 6.946× 10−9
Table 5.2: Adimensional parameters used in the perturbative expressions for the pressure with ξ = 10−11,b = 10−8 and polytropic index n = 3.
And so we reach the end of this work. The equations, although quite complex in form, present
themselves in a more obvious way through this numerical analysis. Through the employment of the LE
equations and some dimensional analysis, we were able to separate the contributions of the Bumblebee
potential from the eld itself. This in turn allowed us to study their individual contribution on a set
of astrophysical bodies modeled through a polytropic equation of state. The obtained constraints are,
as will be discussed in the next section, a lot more stringent than the ones previously obtained in the
literature, thus giving relevance to the results.
44
6Conclusions
45
6. Conclusions
After a brief review of the relevant literature that brings context to the Bumblebee models, we
treated it as a zeroth order perturbation on a set of stars. We employed a set of stars which mimic,
in terms of order of magnitude, the radii and masses of some known stars. These stars where then
described by the Lane-Emden solution for a spherically symmetric body. Assuming that it follows
the underlying symmetry of the problem, we choose a Bumblebee eld with a radial component only.
Because the impact of the eld is considered as a perturbation, an attempt was made in order to
constrain the parameters of the model in such a way as to only cause a variation on the system of
roughly 1%, following the accuracy of our present modelling of the Sun.
The obtained constraint for the value of the potential driving the Bumblebee eld, ξb2 . 10−23, is
many orders of magnitude more stringent than the previously available bound ξb2 . 10−9, obtained
by resorting to tests of Kepler's law using the orbit of Venus [64]; by assuming that, in the presence
of matter, the Bumblebee eld is not relaxed at its vev, this study has also yielded a constraint on
the strength of the corresponding potential, ξ < 1034√AG, which is a new result for these models.
Although only a quadratic potential was considered in this study, the change of the power n showed
very little eect on the proles of the perturbation, as it s eect comes mostly as multiplicative factor
on the potential itself.
Future renements of this method could clearly include the use of a more accurate model for stellar
structure, as well as following a more thorough numerical analysis procedure, eectively solving the
(dierential) modied eld equations to rst order in the model's parameters. However, we must still
take into account that this should only rene the obtained bounds, having no eect on the order of
magnitude of the eect.
The application of this same methodology to the study of galaxies is also possible, in order to
gain further knowledge on the constraints to the parameters of our model, as well as the possibility of
describing galactic dark matter as a manifestation of the Bumblebee dynamics following analogue
eorts in both scalar eld [66] and vectorial Aether models [61, 69, 70]. In doing so, the addition of
a non-vanishing temporal component for a time evolving Bumblebee eld could also be considered, in
order to provide a results relevant at cosmological scales.
46
1
47
Bibliography
[1] G. Guiomar and J. Páramos, Astrophysical Constraints on the Bumblebee Model (Accepted for
publication in Phys. Rev. D).
[2] O. Bertolami, J. Páramos, and S. G. Turyshev, General theory of relativity: Will it survive the
next decade? Proceedings of the 359th WE-Heraeus Seminar on, 2006.
[3] O. Bertolami and J. Páramos, The experimental status of Special and General Relativity, Chap-
ter 22 of the 2014 Springer Handbook of Spacetime.
[4] S. Reinhardt, G. Saatho, H. Buhr, L. A. Carlson, A. Wolf et al., Test of relativistic time dilation
with fast optical atomic clocks at dierent velocities, Nature Phys., vol. 3, pp. 861864, 2007.
[5] S. Herrmann, A. Senger, K. Möhle, M. Nagel, E. V. Kovalchuk, and A. Peters, Rotating optical
cavity experiment testing Lorentz invariance at the 10−17 level, Phys. Rev. D, vol. 80, no. 10, p.
105011, Nov. 2009.
[6] S. Rajagopal and A. Kumar, Ostrogradsky instability and Born-Infeld modied cosmology in
Palatini formalism, preprint available on arXiv:1303.6026.
[7] S. G. Turyshev, J. G. Williams, J. Nordtvedt, Kenneth, M. Shao, and J. Murphy, Thomas W.,
35 years of testing relativistic gravity: Where do we go from here? Lect.Notes Phys., vol. 648,
pp. 311330, 2004.
[8] Y. Su, B. R. Heckel, E. Adelberger, J. Gundlach, M. Harris et al., New tests of the universality
of free fall, Phys. Rev.., vol. D50, pp. 36143636, 1994.
[9] R. V. Pound and G. A. Rebka, Gravitational red-shift in nuclear resonance, Phys. Rev. Lett.,
vol. 3, pp. 439441, Nov 1959.
[10] V. A. Kostelecky, Gravity, Lorentz violation, and the standard model, Phys. Rev.., vol. D69, p.
105009, 2004.
[11] , CPT and Lorentz Symmetry: Proceedings of the Fifth Meeting, Living Rev. Rel., vol. 9,
p. 3, 2006.
[12] V. A. Kostelecky and N. Russell, Data Tables for Lorentz and CPT Violation, Rev.Mod.Phys.,
vol. 83, p. 11, 2011.
48
Bibliography
[13] C. M. Will, The Confrontation between general relativity and experiment, Living Rev.Rel.,
vol. 9, p. 3, 2006.
[14] J. Bell, A Tighter constraint on postNewtonian gravity using millisecond pulsars, Astrophys.J.,
vol. 462, p. 287, 1996.
[15] J. Abraham et al., Observation of the suppression of the ux of cosmic rays above 4× 1019eV,
Phys. Rev..Lett., vol. 101, p. 061101, 2008.
[16] M. Galaverni and G. Sigl, Lorentz Violation in the Photon Sector and Ultra-High Energy Cosmic
Rays, Phys. Rev..Lett., vol. 100, p. 021102, 2008.
[17] G. Rubtsov, P. Satunin, and S. Sibiryakov, Prospective constraints on Lorentz violation from
UHE photon detection, 2013.
[18] R. Abbasi et al., Observation of the ankle and evidence for a high-energy break in the cosmic
ray spectrum, Phys.Lett., vol. B619, pp. 271280, 2005.
[19] R. Vessot, M. Levine, E. Mattison, E. Blomberg, T. Homan et al., Test of Relativistic Gravita-
tion with a Space-Borne Hydrogen Maser, Phys. Rev..Lett., vol. 45, pp. 20812084, 1980.
[20] D. Mattingly, Modern tests of Lorentz invariance, Living Rev.Rel., vol. 8, p. 5, 2005.
[21] O. Bertolami and J. Páramos, Do f(r) theories matter? Phys. Rev. D, vol. 77, p. 084018, Apr
2008.
[22] S. Chandrasekhar, An Introduction to the Study of Stellar Structure, Dover Books, 2010.
[23] P. Curie, Sur la symétrie dans les phénoménes physiques, symétrie d'un champ électrique et d'un
champ magnétique, J. Phys. Theor. Appl., vol. 1, pp. 393415, 1894.
[24] M. P. Haugan and C. Lammerzahl, Principles of equivalence: their role in gravitation physics
and experiments that test them, Lect.Notes Phys., vol. 562, pp. 195212, 2001.
[25] C. N. Yang and R. L. Mills, Conservation of isotopic spin and isotopic gauge invariance, Phys.
Rev., vol. 96, pp. 191195, Oct 1954.
[26] J. H. Christenson, J. W. Cronin, V. L. Fitch, and R. Turlay, Evidence for the 2π decay of the k02
meson, Phys. Rev. Lett., vol. 13, pp. 138140, Jul 1964.
[27] J. Schwinger, On gauge invariance and vacuum polarization, Phys. Rev., vol. 82, pp. 664679,
Jun 1951.
[28] J. P. e. a. Lees, Observation of time-reversal violation in the B0 meson system, Phys. Rev. Lett.,
vol. 109, p. 211801, Nov 2012.
[29] J. Goldstone, Field Theories with Superconductor Solutions, Nuovo Cim., vol. 19, pp. 154164,
1961.
49
Bibliography
[30] T. Bertschinger, N. A. Flowers, and J. D. Tasson, Observer and Particle Transformations and
Newton's Laws, 2013.
[31] V. A. Kostelecky and S. Samuel, Spontaneous Breaking of Lorentz Symmetry in String Theory,
Phys. Rev.., vol. D39, p. 683, 1989.
[32] T. G. Zlosnik, P. G. Ferreira, and G. D. Starkman, Modifying gravity with the aether: An
alternative to dark matter, Physical Review D, vol. 75, no. 4, p. 044017, Feb. 2007.
[33] A. Tartaglia and N. Radicella, Vector eld theories in cosmology, Physical Review D, vol. 76,
no. 8, p. 083501, Oct. 2007.
[34] R. Bluhm, Eects of Spontaneous Lorentz Violation in Gravity, PoS, vol. QG-PH, no. June, p.
009, 2007.
[35] C. Armendariz-Picon and A. Diez-Tejedor, Aether Unleashed, JCAP, vol. 0912, p. 018, 2009.
[36] K. J. Nordtvedt and C. M. Will, Conservation Laws and Preferred Frames in Relativistic Gravity.
II. Experimental Evidence to Rule Out Preferred-Frame Theories of Gravity, Astrophys.J., vol.
177, pp. 775792, 1972.
[37] D. Griths, Introduction to Elementary Particles, ser. Physics textbook. Wiley, 2008.
[38] J. H. Oort, The force exerted by the stellar system in the direction perpindicular to the galactic
plane and some related problems, Bulletin of the Astronomical Institutes of the Netherlands, vol.
249, no. 4, 1932.
[39] F. Zwicky, On the masses of nebulae and clusters of nebulae, Ap.J., vol. 217-246, no. 86, 1937.
[40] W. T. Rubin, Ford, Rotational properties of 21 SC galaxies with a large range of luminosities
and radii, from NGC 4605 R = 4kpc to UGC 2885 R = 122 kpc, Ap.J., vol. 238, Jun. 1980.
[41] A. Vikhlinin et. al, Chandra Sample of Nearby Relaxed Galaxy Clusters: Mass, Gas Fraction,
and Mass-Temperature Relation, Ap.J., vol. 640, pp. 691709, Apr. 2006.
[42] C. et. al, A Direct Empirical Proof of the Existence of Dark Matter, Ap.J., vol. 648, pp. L109
L113, Sep. 2006.
[43] D. Paracz, J.-P. Kneib, J. Richard, A. Morandi, M. Limousin et al., The Bullet Cluster revisited:
New results from new constraints and improved strong lensing modeling technique, 2012.
[44] P. C. XVI, Planck 2013 results: Cosmological parameters (p11), to appear in Astronomy &
Astrophysics, keywords = Astrophysics, year = 2013.
[45] G. Bertone, D. Hooper, and J. Silk, Particle dark matter: Evidence, candidates and constraints,
Phys.Rept., vol. 405, pp. 279390, 2005.
[46] R. D. Peccei and H. R. Quinn, CP conservation in the presence of pseudoparticles, Physical
Review Letters, vol. 38, pp. 14401443, Jun. 1977.
50
Bibliography
[47] L. D. Duy, P. Sikivie, D. Tanner, S. J. Asztalos, C. Hagmann et al., A high resolution search
for dark-matter axions, Phys. Rev.., vol. D74, p. 012006, 2006.
[48] Z. Ahmed et al., Results from a Low-Energy Analysis of the CDMS II Germanium Data, Phys.
Rev..Lett., vol. 106, p. 131302, 2011.
[49] P. J. Fox, G. Jung, P. Sorensen, and N. Weiner, Dark Matter in Light of LUX, 2013.
[50] J. L. Feng, Dark Matter Candidates from Particle Physics and Methods of Detection,
Ann.Rev.Astron.Astrophys., vol. 48, pp. 495545, 2010.
[51] L. Bergstrom, Dark Matter Evidence, Particle Physics Candidates and Detection Methods,
Annalen Phys., vol. 524, pp. 479496, 2012.
[52] T. Bernal, S. Capozziello, G. Cristofano, and M. De Laurentis, MOND's acceleration scale as a
fundamental quantity, Mod.Phys.Lett., vol. A26, pp. 26772687, 2011.
[53] R. H. Sanders and S. S. McGaugh, Modied Newtonian dynamics as an alternative to dark
matter, Ann.Rev.Astron.Astrophys., vol. 40, pp. 263317, 2002.
[54] J. D. Bekenstein, Relativistic gravitation theory for the MOND paradigm, Phys. Rev.., vol. D70,
p. 083509, 2004.
[55] O. Bertolami and J. Páramos, Current tests of alternative gravity Theories: The Modied New-
tonian Dynamics case, Proceedings of the 2006, 36th COSPAR Scientic Assembly.
[56] R. Sanders, Resolving the virial discrepancy in clusters of galaxies with modied newtonian
dynamics, 1998.
[57] G. W. Angus, B. Famaey, and H. Zhao, Can MOND take a bullet? Analytical comparisons of
three versions of MOND beyond spherical symmetry, Mon.Not.Roy.Astron.Soc., vol. 371, p. 138,
2006.
[58] F. Combes, Properties of dark matter haloes, New Astron.Rev., vol. 46, pp. 755766, 2002.
[59] S. M. Carroll and E. A. Lim, Lorentz-violating vector elds slow the universe down, Phys. Rev..,
vol. D70, p. 123525, 2004.
[60] C. Gao, Y. Gong, X. Wang, and X. Chen, Cosmological models with Lagrange Multiplier Field,
Phys.Lett., vol. B702, pp. 107113, 2011.
[61] T. Jacobson, Einstein-aether gravity: A Status report, PoS, vol. QG-PH, p. 020, 2007.
[62] S. Kanno and J. Soda, Lorentz Violating Ination, Phys. Rev.., vol. D74, p. 063505, 2006.
[63] Q. G. Bailey and V. A. Kostelecky, Signals for Lorentz violation in post-Newtonian gravity,
Phys. Rev.., vol. D74, p. 045001, 2006.
51
Bibliography
[64] O. Bertolami and J. Páramos, The Flight of the bumblebee: Vacuum solutions of a gravity model
with vector-induced spontaneous Lorentz symmetry breaking, Phys. Rev.., vol. D72, p. 044001,
2005.
[65] M. D. Seifert, Generalized bumblebee models and Lorentz-violating electrodynamics, Phys.Rev.,
vol. D81, p. 065010, 2010.
[66] O. Bertolami and J. Páramos, Astrophysical constraints on scalar eld models, Phys. Rev. D,
vol. 71, p. 023521, Jan 2005.
[67] O. Bertolami, J. Páramos, and P. Santos, Astrophysical constrains on Ungravity inspired models,
Phys. Rev., vol. D 80, p. 022001, 2009.
[68] J. Casanellas, P. Pani, I. Lopes, and V. Cardoso, Testing alternative theories of gravity using the
Sun, Astrophys.J., vol. 745, p. 15, 2012.
[69] J. D. Barrow, Some Inationary Einstein-Aether Cosmologies, Phys. Rev.., vol. D85, p. 047503,
2012.
[70] W. Donnelly and T. Jacobson, Coupling the inaton to an expanding aether, Phys. Rev.., vol.
D82, p. 064032, 2010.
52