Topics in Bouncing Cosmology PDF

Topics in Bouncing Cosmology
Jerome Quintin
Master of Science
Physics Department
McGill University
3600 rue University, Montréal, Québec, H3A 2T8, Canada
2015-04-15
A thesis submitted to McGill University in partial fulfillment of the requirements of

the degree of Master of Science
Jerome
c Quintin 2015.
ACKNOWLEDGEMENTS
I would first like to thank my fellow graduate students for exciting discussions. In
particular, I thank Elisa G. M. Ferreira, Marc-Antoine Fiset, and Evan McDonough
for their patience and invaluable help. I then thank Yi-Fu Cai for his tremendous
generosity and time, and my other collaborators, Elcio Abdalla, Fang Chen, Em-
manuel N. Saridakis, Bin Wang, and Edward Wilson-Ewing, with whom it was a
great pleasure to do research. I am also thankful to my parents, and family, for their
constant support and encouragement throughout the process of my degree.
I would like to express my deep and sincere gratitude to my supervisor Pro-

fessor Robert H. Brandenberger for his generosity to transmit his knowledge. This
work benefited greatly from his patience, his guidance, and his tremendous help.
Finally, I thank McGill University, the Natural Sciences and Engineering Re-
search Council (NSERC) of Canada, and the Fonds de recherche du Québec - Nature
et technologies (FRQNT) for providing funding during the completion of this degree.
ii
Contributions of Author for the Papers included in this Thesis
• Y. F. Cai, J. Quintin, E. N. Saridakis and E. Wilson-Ewing, Nonsingular bounc-
ing cosmologies in light of BICEP2, JCAP 1407, 033 (2014) [arXiv:1404.4364
[astro-ph.CO]].
For this paper, I mainly focused on the sections on matter bounce cosmology
where I carried out all of the analytical calculations and numerical computa-
tions, and where I made significant contributions to the writing. I also largely
contributed to the writing of the introduction and conclusion. Y.-F. Cai made
significant contributions to the writing of the introduction, the matter bounce
cosmology section, and the conclusion. He also wrote most of the matter bounce
in Hořava-Lifshitz gravity section. E. N. Saridakis carried out all of the cal-
culations of the f (T ) matter bounce cosmology section and wrote all of it.
E. Wilson-Ewing carried out all of the calculations of the loop quantum cos-
mology section and wrote all of it. He also contributed to the writing of the
conclusion.
• J. Quintin, Y. F. Cai and R. H. Brandenberger, Matter creation in a nonsingu-

lar bouncing cosmology, Phys. Rev. D 90, no. 6, 063507 (2014) [arXiv:1406.6049
[gr-qc]].
For this paper, I carried out all of the calculations and contributed to the
writing of most of the text (around 85%). Y. F. Cai contributed to the writing
of the gravitational particle production section and assisted me through the
calculations. R. H. Brandenberger contributed to the writing of the introduc-
tion, the conclusion, and the rest of the text. He also assisted me through the
calculations.
Papers not included in this Thesis

• E. Abdalla, E. G. M. Ferreira, J. Quintin and B. Wang, New Evidence for
Interacting Dark Energy from BOSS, arXiv:1412.2777 [astro-ph.CO].
• Y. F. Cai, F. Chen, E. G. M. Ferreira and J. Quintin, A new model of axion

monodromy inflation and its cosmological implications, arXiv:1412.4298 [hep-
th].
iii
• S. F. Bramberger, R. H. Brandenberger, P. Jreidini and J. Quintin, Cosmic
String Loops as the Seeds of Super-Massive Black Holes, arXiv:1503.02317
[astro-ph.CO].
iv
ABSTRACT
This thesis concerns two questions in bouncing cosmology, as addressed in the

papers arXiv:1404.4364 and arXiv:1406.6049.
In the first part of this thesis, we explore the possibility for several models
of bouncing cosmology to produce a tensor-to-scalar ratio of the order of 0.2. We
start by analyzing the matter bounce curvaton scenario and the matter-Ekpyrotic
scenario, two models which use matter fields violating the Null Energy Condition
at high energy. We then analyze the Hořava-Lifshitz bounce, the f (T ) bounce, and
the Loop Quantum Cosmology bounce, three models which use modified gravity to
avoid the singularity. We generally find that these matter bounce scenarios produce
large amounts of primordial gravitational waves, but the tensor-to-scalar ratio can
be suppressed for some models, for instance via the curvaton mechanism.
In the second part of this thesis, we explore two possible reheating mechanisms
in the context of the matter-Ekpyrotic bounce, namely gravitational particle produc-
tion and particle creation via the coupling of the bounce field with the matter field.
We find that gravitational particle production, where the background field couples
only to gravity, is sufficient to produce radiation at high enough temperatures after
the bounce. Thus, there seems to be no need to introduce an extra coupling to the
model Lagrangian, contrary to inflation.
v
ABRÉGÉ
Cette thèse concerne deux questions sur la théorie de l’univers rebondissant,

telles qu’adressées dans les articles arXiv:1404.4364 et arXiv:1406.6049.
Dans la première partie de cette thèse, nous explorons la possibilité de produire

un rapport tenseur-sur-scalaire de l’ordre de 0.2 dans le cadre de plusieurs modèles
d’univers rebondissant. Nous commençons par analyser le modèle d’univers rebondis-
sant composé d’un champ scalaire de type curvaton et le modèle composé d’un champ
de matière et de type Ekpyrotic. Ces deux modèles utilisent des champs qui brisent
la condition d’énergie de genre lumière à haute énergie. Nous analysons ensuite les
modèles d’univers rebondissant utilisant la théorie de la gravité de Hořava-Lifshitz,
puis ensuite la théorie modifiée de la gravitation de type f (T ), et la théorie de la
cosmologie quantique à boucles. Ces trois modèles utilisent des théories modifiées de
la gravitation pour éviter la singularité. Nous concluons qu’en général ces scénarios
produisent des ondes gravitationnelles primordiales en grande quantité, mais le rap-
port tenseur-sur-scalaire peut être réduit pour certains modèles, par exemple grâce
au mécanisme de curvaton.
Dans la deuxième partie de cette thèse, nous explorons deux méchanismes de

réchauffement pour le modèle d’univers rebondissant composé d’un champ de matière
et de type Ekpyrotic : la production de particules dans un champ gravitationel et
grâce au couplement du champ de rebondissement avec le champ de matière. Nous
vi
concluons que la production de particules dans un champ gravitationel, où le champ
dominant interagit uniquement avec la gravité, produit suffisamment de radiation
à haute température après le rebondissement. Ainsi, il ne semble pas nécessaire
d’introduire des interactions supplémentaires au Lagrangien du modèle contraire-
ment à la théorie de l’inflation.
vii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
ABRÉGÉ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Review of Bouncing Cosmology . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Bouncing cosmology as a solution to the problems of standard Big
Bang cosmology . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 The flatness problem . . . . . . . . . . . . . . . . . . . . . 5
2.1.2 The horizon problem . . . . . . . . . . . . . . . . . . . . . 6
2.1.3 The formation of structure problem . . . . . . . . . . . . . 6
2.2 A short review of the theory of cosmological perturbations . . . . 7
2.3 The generation of a scale-invariant power spectrum of curvature
perturbations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.4 The models of nonsingular bouncing cosmology . . . . . . . . . . 13
2.5 The matter-Ekpyrotic bounce . . . . . . . . . . . . . . . . . . . . 14
2.6 The problems of the matter-Ekpyrotic bounce . . . . . . . . . . . 17
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Nonsingular Bouncing Cosmologies in Light of BICEP2 . . . . . . . . . . 26
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2 Matter Bounce Cosmology . . . . . . . . . . . . . . . . . . . . . . 28
3.2.1 The matter bounce curvaton model . . . . . . . . . . . . . 28
3.2.2 New matter bounce cosmology . . . . . . . . . . . . . . . . 31
viii
3.3 Implications for modified-gravity bouncing cosmology . . . . . . . 38
3.3.1 Matter bounce in Hořava-Lifshitz gravity . . . . . . . . . . 38
3.3.2 The f (T ) matter bounce cosmology . . . . . . . . . . . . . 39
3.3.3 Loop quantum cosmology . . . . . . . . . . . . . . . . . . . 40
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4 Matter Creation in a Nonsingular Bouncing Cosmology . . . . . . . . . . 52
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.2 The Two-Field Matter Bounce Model . . . . . . . . . . . . . . . . 55
4.3 Gravitational particle production throughout the cosmic evolution 59
4.3.1 Dynamics of the variable χk . . . . . . . . . . . . . . . . . 62
4.3.2 Particle production . . . . . . . . . . . . . . . . . . . . . . 67
4.3.3 Reheating time and reheating temperature . . . . . . . . . 73
4.4 Backreaction of Thermal Particles from Cosmic Fluctuations . . . 75
4.5 Particle Production Through Direct Interactions . . . . . . . . . . 76
4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Solving for the Bogoliubov coefficients in the bounce phase . . . . . . . . . . . 85
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5 Conclusions and Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
ix
LIST OF FIGURES
Figure page
3–1 Constraints on the mass parameter m and the slope parameter Υ of
the bounce phase from the measurements of Planck and BICEP2 in
the matter bounce curvaton scenario. The blue bands show the 1σ
and 2σ confidence intervals of the tensor-to-scalar ratio and the red
bands show the confidence intervals of the amplitude of the power
spectrum of curvature perturbations. . . . . . . . . . . . . . . . . . 32
3–2 Constraints on the dimensionless duration parameter tB+ /T and the
slope parameter Υ of the bounce phase from the measurements of
Planck and BICEP2 in the new matter bounce cosmology. The
value of the Hubble parameter at the beginning of the ekpyrotic
phase is fixed to be HE /Mp = 10−7 . As in Fig. 3–1, the blue and red
bands show the confidence intervals of the tensor-to-scalar ratio and
of the amplitude of the power spectrum of curvature perturbations,
respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3–3 Constraints on the Hubble parameter at the beginning of the ekpyrotic
phase HE and the slope parameter Υ of the bounce phase from the
measurements of Planck and BICEP2 in the new matter bounce
cosmology. The dimensionless bounce time duration is fixed to be
tB+ /T = 1.86. As in Figs. 3–1 and 3–2, the blue and red bands show
the confidence intervals of r and of the amplitude of Pζ , respectively. 37
x
4–1 Space-time sketch in the two-field matter bounce scenario. The
comoving length and conformal time are labeled by x and η,
respectively. The solid blue curve shows H−1 , the comoving Hubble
radius. The different phase transition times are labeled on the
vertical axis. The shaded areas show the regions of integration
for the respective phases (the Ekpyrotic and bounce phases are
shown). The comoving wavelengths of the fluctuating modes that
we integrate originate in these shaded areas. The green line shows
the comoving wavelength of such a fluctuating mode that forms in
the Ekpyrotic phase of contraction. We see that this mode reenters
the Hubble horizon at later times, after ηB+ . . . . . . . . . . . . . . 58
4–2 Plot of the ratio of the energy density of particle production to the
background energy density as a function of the coupling constant.
The slope of the Hubble parameter and the physical duration of the
bounce are taken to be Υ = 5.5 × 10−7 Mp2 and tB+ = 2.1 × 103 Mp−1 ,
respectively. The blue curve shows the analytic upper bound found
in Eq. (4.84) and the black curve was obtained by solving the EoM
and the integral for the energy density numerically. The dashed
red line is the contribution from Parker particle production that we
found earlier [see Eq. (4.54)]. . . . . . . . . . . . . . . . . . . . . . 82
xi
Chapter 1
Introduction
The standard Big Bang model of cosmology comes from Einstein’s general the-
ory of relativity, which describes gravity and spacetime, together with the assump-
tions that the universe looks the same in every direction and that it is spatially
homogeneous. The resulting standard model of cosmology successfully describes the
universe for a number of observations, but it also has several important problems
(see, e.g., [1]). In particular, it cannot explain why the universe appears to be al-
most perfectly flat and homogeneous, but more importantly, it cannot explain the
formation of structures in our universe.
Many of these problems can be solved with the introduction of a short but
dramatic phase of exponential expansion of the universe in its very early times. This
leads to the theory of inflationary cosmology [2–5] (see, e.g., [6,7] for reviews), which
has become the paradigm of early universe cosmology because of its predictions [8]
and since there is now much evidence to support it (see [9] for a review of the
observational status). In fact, some authors such as Guth et al. [10] claim that
recent observations coming from telescopes such as the Planck experiment [11, 12]
have reinforced inflation as the most plausible theory of early universe cosmology.
The theory of inflation, however, is far from being a perfect theory. In fact,
it suffers from many serious conceptual problems (see, e.g., [13, 14] and references
therein). For instance, inflation cannot resolve the singularity problem of standard
Big Bang cosmology, which says that the universe had an infinite temperature and
an infinite energy density at the initial time of the Big Bang. This suggests that one
should search for alternatives to inflation alongside the research that is being carried
out on inflation.
Two promising alternatives to inflation are the theory of string gas cosmology
and the matter bounce scenario (see, e.g., [15–20] for recent discussions and reviews).
1
This thesis will focus on the latter, where quantum fluctuations originate in a matter-
dominated contracting phase, and evolve through a nonsingular bounce where new
physics may appear. The background cosmology is followed by the phases of standard
Big Bang cosmology.
This thesis first starts, in Chapter 2, by reviewing the theory of bouncing cos-
mology. We begin by explaining how bouncing cosmology solves the problems of
standard Big Bang cosmology. We then give a short review of the theory of cosmo-
logical perturbations before showing how one can generate a scale-invariant power
spectrum of curvature perturbations. We then describe the different models of bounc-
ing cosmology, putting more emphasis on the ghost condensate method. We move
on to introduce the matter-Ekpyrotic bounce. In the end of this chapter, we sum-
marize the current status of the theory with regards to observations, and motivate
the two articles presented in this thesis. First, Chapter 3 presents the article by Cai
et al. [21], which discusses the possibility of several models of bouncing cosmology
to produce a tensor-to-scalar ratio of the order of 0.2. Second, Chapter 4 presents
the article by Quintin et al. [22], which explores different reheating mechanisms in
the context of the matter-Ekpyrotic nonsingular bouncing cosmology. Finally, we
conclude and give an outlook in Chapter 5.
2
Bibliography
[1] R. H. Brandenberger, Inflationary cosmology: Progress and problems, hep-
ph/9910410.
[2] A. H. Guth, The Inflationary Universe: A Possible Solution to the Horizon and
Flatness Problems, Phys. Rev. D 23, 347 (1981).
[3] R. Brout, F. Englert and E. Gunzig, The Creation of the Universe as a Quantum
Phenomenon, Annals Phys. 115, 78 (1978).
[4] K. Sato, First Order Phase Transition of a Vacuum and Expansion of the Uni-
verse, Mon. Not. Roy. Astron. Soc. 195, 467 (1981).
[5] A. A. Starobinsky, A New Type of Isotropic Cosmological Models Without Sin-
gularity, Phys. Lett. B 91, 99 (1980).
[6] A. D. Linde, Inflationary Cosmology, Lect. Notes Phys. 738, 1 (2008)
[arXiv:0705.0164 [hep-th]].
[7] D. Baumann, TASI Lectures on Inflation, arXiv:0907.5424 [hep-th].
[8] V. F. Mukhanov and G. V. Chibisov, Quantum Fluctuation and Nonsingular
Universe, JETP Lett. 33, 532 (1981) [Pisma Zh. Eksp. Teor. Fiz. 33, 549 (1981)].
[9] J. Martin, The Observational Status of Cosmic Inflation after Planck,
arXiv:1502.05733 [astro-ph.CO].
[10] A. H. Guth, D. I. Kaiser and Y. Nomura, Inflationary paradigm after Planck
2013, Phys. Lett. B 733, 112 (2014) [arXiv:1312.7619 [astro-ph.CO]].
[11] P. A. R. Ade et al. [Planck Collaboration], Planck 2013 results. XXII. Con-
straints on inflation, Astron. Astrophys. 571, A22 (2014) [arXiv:1303.5082 [astro-
ph.CO]].
[12] P. A. R. Ade et al. [Planck Collaboration], Planck 2015. XX. Constraints on
inflation, arXiv:1502.02114 [astro-ph.CO].
3
4
[13] R. H. Brandenberger, Cosmology of the Very Early Universe, AIP Conf. Proc.
1268, 3 (2010) [arXiv:1003.1745 [hep-th]].
[14] R. Brandenberger, Do we have a Theory of Early Universe Cosmology?, Stud.
Hist. Philos. Mod. Phys. 46, 109 (2014) [arXiv:1204.6108 [astro-ph.CO]].
[15] R. H. Brandenberger, String Gas Cosmology: Progress and Problems, Class.
Quant. Grav. 28, 204005 (2011) [arXiv:1105.3247 [hep-th]].
[16] R. H. Brandenberger, The Matter Bounce Alternative to Inflationary Cosmology,
[17] R. H. Brandenberger, Alternatives to the inflationary paradigm of structure for-
mation, Int. J. Mod. Phys. Conf. Ser. 01, 67 (2011) [arXiv:0902.4731 [hep-th]].
[18] R. H. Brandenberger, Unconventional Cosmology, Lect. Notes Phys. 863, 333
(2013) [arXiv:1203.6698 [astro-ph.CO]].
[19] M. Novello and S. E. P. Bergliaffa, Bouncing Cosmologies, Phys. Rept. 463, 127
(2008) [arXiv:0802.1634 [astro-ph]].
[20] D. Battefeld and P. Peter, A Critical Review of Classical Bouncing Cosmologies,
[21] Y. F. Cai, J. Quintin, E. N. Saridakis and E. Wilson-Ewing, Nonsingular bounc-
[astro-ph.CO]].
[22] J. Quintin, Y. F. Cai and R. H. Brandenberger, Matter creation in a nonsingular
bouncing cosmology, Phys. Rev. D 90, no. 6, 063507 (2014) [arXiv:1406.6049 [gr-
qc]].
Chapter 2
Review of Bouncing Cosmology
2.1 Bouncing cosmology as a solution to the problems of standard Big
Bang cosmology
As mentioned in the introduction, the standard model of cosmology, known as
the hot Big Bang model, comes from Einstein’s general theory of relativity and as-
sumes that our universe is spatially homogeneous and isotropic on very large scales.
From the point of view of modern cosmology, this model cannot explain the origin of
the structures of our universe unless the initial conditions at the time of the Big Bang
are extremely fine-tuned. For this reason, a successful theory of the very early uni-
verse must at least solve the main three problems of standard Big Bang cosmology.
Let us describe these problems and how they are resolved in the case of bouncing
cosmology.
2.1.1 The flatness problem

From the Friedmann equations and the continuity equation in standard cos-
mology, one can see that a flat universe in an unstable fixed point in an expanding
universe with equation of state parameter w > −1/3 (see, e.g., [1]; here, w ≡ P/ρ,
where P denotes pressure and ρ denotes energy density, so for example, w = 0 for
dust and w = 1/3 for radiation). Observations indicate that our universe is almost
perfectly flat today (see [2] for the latest results), and thus, we have to require a
universe excessively close to flatness in the early times, which would be an extremely
fine-tuned initial condition. This issue is what we call the flatness problem.
In bouncing cosmology, the history of the universe starts in the infinite past
and the universe is infinitely large at that moment. In other words, time runs from
−∞ to +∞, and there is no beginning. For t < 0, the universe is in a contracting
phase, followed by the bounce phase, at which point there is a transition between
5
contraction and expansion. Then, the universe enters an expanding phase (for t > 0)
just as in standard cosmology.
We said earlier that a flat universe was an unstable fixed point in an expanding
universe. It turns out that it is a stable fixed point in a contracting universe (still
with w > −1/3). Thus, whatever the initial conditions are as t → −∞ in a bouncing
universe, the universe will be asymptotically flat as t → 0, and consequently, very
close to perfect flatness today.
2.1.2 The horizon problem

The Cosmic Microwave Background (CMB) is observed to be homogeneous to
better than 10−4 . In order to attain such homogeneity, within the observed region
of the CMB, the causal past of two photons must have intersected to be in thermal
equilibrium. Thus, the size of the comoving light cone at the time of the CMB
(called the time of recombination, trec ) should be of the same order as the size of the
comoving region over which the CMB is observed. Let us denote the former lp (trec )
and the latter lf (trec ). A simple calculation shows that (see, e.g., [3])
1/3
lp (trec ) t0
' −1 , (2.1)
lf (trec ) trec
where t0 denotes the time today and obviously t0 trec . Therefore, we have that
lp (trec ) lf (trec ), and so it appears that the universe is too small at the time of
recombination to match the observations of the CMB. This is what we call the
horizon problem.
In bouncing cosmology, there is a long period of time before the “Big Bang” for
photons to interact and get in thermal equilibrium. In fact, if the contracting period
before the bounce is as long as the time between the bounce and today [such that
lf (trec ) & lp (trec )], it is obvious that there is no horizon problem.
2.1.3 The formation of structure problem

We know that the formation of large scale structures is due to small density
perturbations in the early universe. On one hand, if these initial perturbations
6
were produced after the time of matter-radiation equality (teq ), then gravity would
generally be too weak to produce the observed galaxy clusters. On the other hand, if
these initial perturbations were produced before teq , then the observed correlations of
galaxy clusters could not be explained by a causal mechanism because it is known that
the scale of these correlations is larger than the comoving horizon at teq . Therefore,
standard cosmology cannot explain the observed correlations, hence the formation
of structure problem.
In bouncing cosmology, the initial perturbations originate in the contracting
phase before the bounce on sub-Hubble scales. For the same reason that bouncing
cosmology is free of the horizon problem, there is a causal mechanism for these per-
turbations to evolve into the observed large scale structures of our universe. In the
following sections, we will see that these initial perturbations grow in a contracting
universe, and we will see that their resulting spectrum is scale-invariant if they have
a quantum origin and if they exit the Hubble radius in a matter-dominated contract-
ing phase. But first, we must review the theory of cosmological perturbations.
2.2 A short review of the theory of cosmological perturbations

The general idea of the theory of cosmological perturbations (see, e.g., [4–7] for
reviews and [8,9] for textbook descriptions) is to introduce small linear perturbations
about some background cosmology. In most cases, the background metric is taken
to be the Friedmann-Lemaître-Robertson-Walker (FLRW) metric,
ds2 = gµν
(0)
dxµ dxν = a2 (η)(dη 2 − dx2 ) , (2.2)
where gµν is the metric tensor, a is the scale factor, η represents conformal time,
and x is the comoving spatial coordinate. Conformal time is defined in terms of the
cosmic physical time t as follows:
dη ≡ a−1 dt . (2.3)
7
The superscript (0) indicates background quantities. The action for matter and
gravity is taken to be
√
Z
4 1 2
S = d x −g − Mp R + Lm , (2.4)
2
where we define the reduced Planck mass to be Mp2 ≡ 1/8πGN . Here, R is the
Ricci scalar, and Lm is the matter Lagrangian density. We will assume that the
Lagrangian depends on one scalar field, φ. Introducing small perturbations about
the background amounts to taking
(0)
gµν (η, x) = gµν (η) + δgµν (η, x) , (2.5)
φ(η, x) = φ(0) (η) + δφ(η, x) . (2.6)
Scalar metric fluctuations are generally represented by four degrees of freedom,

" #
2Φ −B ,i
δgµν = a2 . (2.7)
−B,i 2(Ψδij − E,ij )
Vector metric fluctuations have four degrees of freedom as well, but they will be ig-
nored throughout this thesis because they decay in an expanding universe 1 . Finally,
tensor modes have two degrees of freedom,
" #
0 0
δgµν = a2 , (2.8)
0 hij
1
Since we consider a contracting universe in bouncing cosmology, one may worry
that vector modes become important. In fact, these modes would grow in a con-
tracting universe and they would be maximal at the bounce point. However, they
would decay as the universe expands after the bounce. Since we do not measure
perturbations at the time of the Big Bang, we do not observe these modes. One may
worry that the vector modes couple to other modes and leave some imprint in these
modes, but at linear order in perturbation, this does not occur.
8
where hij is traceless and divergence-less.
Upon gauge fixing (see the reviews cited earlier for details), one can work out
the equations of motion for the resulting fluctuation degrees of freedom. We choose
the conformal Newtonian gauge here (E = B = 0), in which case the perturbed
metric for scalar modes becomes
ds2 = a2 (η) [1 + 2Φ(η, x)] dη 2 − [1 − 2Φ(η, x)] dx2 ,

(2.9)
assuming there is no anisotropic stresses in the matter content to linear order (i.e.
Φ = Ψ). Similarly, the perturbed metric for tensor modes is
ds2 = a2 (η) dη 2 − [δij + hij (η, x)] dxi dxj

. (2.10)
Substituting these perturbed metric in the action and expanding up to second order
in perturbed quantities, one finds
z 00 2
Z
(2) 1 4 0 2 2
Sscalar = d x (v ) − (v,i ) + v , (2.11)
2 z
a00 2
Z
(2) 1 4 0 2 2
Stensor = d x (µ ) − (µ,i ) + µ , (2.12)
2 a
where a prime denotes a derivative with respect to conformal time and where we
introduced the variables z, v, and µ. For a canonical scalar field with Lagrangian
1
Lm = g µν ∇µ φ∇µ φ − V (φ) , (2.13)
2
the variable z takes the form
a dφ(0)
z= , (2.14)
H dη
with H ≡ a0 /a. Introducing the curvature perturbation R defined by
H
R≡ δφ + Φ , (2.15)
φ(0)0
9
the variable v, known as the Mukhanov variable, is given by
v = zR . (2.16)
Equivalently for tensor modes, the Mukhanov variable is given by
µ = ah , (2.17)
where h is a polarization state of the tensor perturbation hij .

Varying the perturbed actions S (2) for scalar and tensor modes, one finds the
equations of motion,
z 00
v 00 − ∇2 v − v=0, (2.18)
z
a00
µ00 − ∇2 µ − µ = 0 , (2.19)
a
which are more commonly written in Fourier space as follows:
z 00

00 2
vk + k − vk = 0 , (2.20)
z
a00

00 2
µk + k − µk = 0 . (2.21)
a
Equivalently, the equation of motion (EOM) for the curvature perturbation is
z0
R00k + 2 R0k + k 2 Rk = 0 , (2.22)
z
and the EOM for the tensor perturbation is
a0
h00k + 2 h0k + k 2 hk = 0 . (2.23)
a
From the solutions of these EOMs, one can define the power spectrum of curva-
ture perturbations by
k3 2 k 3 −2
PR (k, η) ≡ |Rk (η)| = z (η)|vk (η)|2 . (2.24)
2π 2 2π 2
10
We give an example of the calculation of the power spectrum of curvature perturba-
tions in the next section. Equivalently, we can define the power spectrum of tensor
perturbations by
k3 2 k 3 −2
Ph (k, η) ≡ |hk (η)| = a (η)|µk (η)|2 . (2.25)
2π 2 2π 2
2.3 The generation of a scale-invariant power spectrum of curvature per-

turbations
Let us describe the background cosmological evolution by the following scale
factor,
a(t) ∼ (−t)p , (2.26)
for some power p and in physical time t. Furthermore, let us assume that the equation
of state for the matter content of the universe does not evolve, in which case one can
show that the variable z introduced in the previous section is proportional to the
scale factor. Thus, one finds that
z 00 a00 p(2p − 1) −2
= = η , (2.27)
z a (p − 1)2
where we used the relation dη = a−1 dt. Then the equation of motion for the scalar
modes as introduced in the previous section becomes

00 2 2 1 −2
vk + k − ν − η vk = 0 , (2.28)
4
where we defined
(1 − 3p)2
ν2 ≡ . (2.29)
4(p − 1)2
We recognize this EOM to be a special form of the Bessel equation, and its solution
is given by
√ √
vk (η) = c1 (k) −ηJ|ν| (−kη) + c2 (k) −ηY|ν| (−kη) , (2.30)
where J|ν| and Y|ν| are the Bessel functions of the first and second kind, respectively.
11
Taking a quantum vacuum of fluctuations as the origin of the large scale struc-
tures that we observe today, we impose the solution for vk to be the Bunch-Davies
vacuum,
1
vk (η) = √ e−ikη , (2.31)
2k
in the limit |kη| → ∞ as the initial conditions. The resulting solution is
√
−πη i(ν+ 1 ) π (1)
vk (η) = e 2 2H
|ν| (−kη) , (2.32)
2
(1)
where H|ν| (−kη) ≡ J|ν| (−kη) + iY|ν| (−kη) is the Hankel function.
The resulting power spectrum for long wavelength curvature fluctuations (i.e. in
the limit kη → 0) is
PR (k) ∼ k 3−2|ν| . (2.33)
Thus, we see that a scale-invariant (i.e. k-independent) power spectrum of curvature

perturbations occurs for |ν| = 3/2, which is the case when p = 2/3 or when p → ∞.
In the first case, we have a ∼ (−t)2/3 , so a matter-dominated contracting background
cosmology can lead to a scale-invariant power spectrum. In the second case, if we
substitute −t for t (which yields the same solution), the scale factor grows as a power
law with an infinite power which is equivalent to having an exponentially growing
scale factor, i.e. a ∼ eHt for some constant H, and so an inflationary background
cosmology can also yield a scale-invariant power spectrum.
The fact that inflationary cosmology could lead to a scale-invariant power spec-
trum of curvature perturbations was known as early as in the late 1970’s and early
1980’s (see, e.g., [10–16]) and the theory of inflationary cosmology has considerably
evolved since then. But it was only much later that the fact that a matter-dominated
contracting universe could also lead to a scale-invariant power spectrum was discov-
ered. It was first noticed in the late 1990’s by Wands [17] and the idea was further
developed in the early 2000’s by Finelli and Brandenberger [18]. In the following
years, the matter bounce scenario and more general bouncing cosmologies have been
investigated. We discuss some of these models in the following section.
12
2.4 The models of nonsingular bouncing cosmology
Since a matter-dominated contracting universe leads to a scale-invariant power
spectrum of curvature perturbations, it is natural to build a model with such a real-
ization. The matter bounce scenario proposes that before the Big Bang, there was
a matter-dominated contracting phase in which vacuum quantum fluctuations origi-
nated. In addition, the model requires that modes with wavelength of cosmological
interest today exited the Hubble radius during the matter-dominated phase.
According to the Penrose-Hawking singularity theorems in the context of Gen-
eral Relativity (GR), any cosmological model contains, inevitably, a singularity
(see [19] and references therein). Therefore, the matter bounce as described above
would unavoidably reach a singularity after the contracting phase. However, this is
only true under certain assumptions. In general, the singularity may be avoided at
high energies if one of the following conditions holds:
• gravity is described by a modified gravity theory, in which GR is only a low-
energy effective description;
• there exists quantum gravity effects at high energies;
• there exists matter which violates the Null Energy Condition (NEC) at high
energies.
It turns out that one can realize a nonsingular matter bounce model for each of
these conditions. It has been realized for a number of different modifications to
gravity (see, e.g., [20–26]), and for popular quantum gravity proposals such as in
String Theory (e.g. [27]), Loop Quantum Cosmology (e.g. [28–33]), and Hořava-
Lifshitz gravity [34,35]. There are also many ways of obtaining a nonsingular bounce
using modified matter which violates the NEC. Examples include quintom matter
[36–41] such as in the Lick-Wick model [42–45], but these models usually suffer from
many instability problems (see, e.g, [46–48]). An alternative consists in realizing a
nonsingular bouncing cosmology with a ghost condensate [49–53].
To describe the idea of the ghost condensate, let us introduce the kinetic variable
X for a scalar field φ:
1
X ≡ g µν ∇µ φ∇ν φ . (2.34)
2
13
We temporarily take the scalar field to have units of length so that X is dimensionless.
If the Lagrangian density has the form
L(φ, X) = (X − C 2 )2 − V (φ) , (2.35)
where V (φ) is some potential function, and C is some dimensionless constant, then
we notice that the kinetic part of the theory has the form −2XC 2 + X 2 , where the
leading term in X appears to have the wrong sign and thus appears to be unstable
against a ghost instability. However, from the form of the above Lagrangian, we
see that there exists a non-trivial minimum at X = C 2 , which is called the ghost
condensate, about which fluctuations have the correct kinetic sign. In realizing a
nonsingular bounce with a ghost condensate and with an appropriate potential, one
can further avoid instabilities such as instabilities against the addition of radiation
and anisotropic stress during the bounce [49]. However, the model is still unstable
against the addition of anisotropic stress during the matter-dominated contracting
phase before the nonsingular bounce. This is known as the BKL instability [54]. The
following section elaborates on this problem and how it can be solved. For a more
thorough review of the different models of nonsingular bouncing cosmology and of
their problems and advantages, we refer the reader to reviews such as [55, 56].
2.5 The matter-Ekpyrotic bounce

It is well-known that the energy density of matter scales as a−3 and that the
energy density of radiation scales as a−4 . Thus, in the context of a contracting
universe with a mixture of matter and radiation, matter dominates at early times
after which there is a radiation-dominated contacting phase. However, the energy
density associated with anisotropic stress scales as a−6 and thus always dominates
over matter and radiation in a contracting universe after some time. The instability
to the growth of anisotropic stress in this context is known as the BKL instability [54].
One solution to the BKL instability resides in the addition of an Ekpyrotic phase.
The initial Ekpyrotic scenario was introduced in [57, 58] (see [59] for a review) from
14
a string theory perspective, and was modeled by a 5-dimensional spacetime consist-
ing of two (3+1)-dimensional branes that act as boundaries of the spacetime, sepa-
rated by a finite dimension. In this setup, gravity lives in the 5-dimensional space-
time, whereas the other fundamental forces and matter live in the (3+1)-dimensional
branes. The distance between the branes acts as a modulus of the theory and the
potential of the resulting scalar field is taken such that there is an attractive force
between the two branes. Because of that, the branes approach one another and this
is called the Ekpyrotic phase of contraction. This phase can be modeled by a scalar
field φ with potential V (φ) = −VE e−cφ , where VE and c are constants. This immedi-
ately leads to an equation of state parameter w ≡ P/ρ > 1 (P denotes pressure and
ρ denotes energy density). Furthermore, if one takes c 1, then the background
cosmology is described by
a(t) = (−t)p , (2.36)

s !
p VE
φ(t) = 2p ln − t , (2.37)
p
where p ≡ 2/c2 1. This describes a contracting universe with equation of state

parameter
2
w= −11 . (2.38)
3p
The fact that w 1 is typically taken as the definition of an Ekpyrotic phase of
contraction.
The interesting fact about an Ekpyrotic phase of contraction is that the energy
density scales as a−2/p , where 2/p is usually much greater than 6. Therefore, the
Ekpyrotic field is dominant over anisotropic stress, and thus this model can avoid
the BKL instability [60]. For this reason, Cai et al. [61] had the idea of adding
an Ekpyrotic phase to the ghost condensate nonsingular matter bounce scenario
introduced in the previous section, and it was later shown to be free of the BKL
instability [62]. The model was further developed in [63].
15
Specifically, the Lagrangian for this model is taken to be
L = K(φ, X) + G(φ, X)φ , (2.39)
where we introduced X in Eq. (2.34) and where ≡ g µν ∇µ ∇ν is the d’Alembertian.

To obtain the Ekpyrotic phase of contraction and the ghost condensate bounce, one
takes2
K(φ, X) = Mp2 [1 − g(φ)]X + βX 2 − V (φ) , (2.40)
where β is a positive constant. When g(φ) > 1, we see that the leading kinetic term
has the “wrong sign” and a ghost condensate is formed as described in the previous
section. Thus, one takes g(φ) to be of the form
2g0
g(φ) = q
2
q
2
, (2.41)
− φ bg φ
e p
+e p
where g0 > 1, p, and bg are constants, so that when φ is close to 0, g(φ) > 1 leads to
a ghost condensate, and when |φ| 1, g(φ) is very small. Similarly, one takes the
potential to be of the form
2V0
V (φ) = − q q , (2.42)
− 2q φ bV 2
φ
e +e q
where V0 , q, and bV are constants, so that when φ approaches 0 from negative values,
the potential is of the form introduced above to obtain an Ekpyrotic phase of con-
traction. Finally, one takes G(φ, X) = γX for some positive constant γ to ensure the
stability of the model. Further details about this model are given in the introduction
to Chapter 4.
2
We adopt the convention that φ is dimensionless throughout the rest of this
thesis.
16
2.6 The problems of the matter-Ekpyrotic bounce
So far, we have described how to obtain a healthy nonsingular matter bounce
from a theoretical perspective, but a successful theory of the early universe should
match observations. It is useful to define the scalar spectral index ns by
d ln PR
ns − 1 ≡ , (2.43)
d ln k
where we recall the definition of PR in Eq. (2.24). One can also define the tensor-
to-scalar ratio by
Ph
r≡ , (2.44)
PR
where we recall the definition of Ph in Eq. (2.25). The original matter bounce
model predicted a scale-invariant power spectrum of curvature perturbation, or in
other words, ns = 1. However, observations indicate a red tilt of the scalar power
spectrum, and rule out ns = 1 at more than 5σ. More specifically, the latest results
give [64]
ns = 0.9645 ± 0.0049 (68% CL) . (2.45)
Thus, a successful bouncing cosmology would need to postdict such a red tilt. This
may be a problem for the matter-Ekpyrotic model as found in [63], but may be
achievable in other bouncing models (see, e.g., [65, 66]). We do not address this
problem in this thesis, but it certainly deserves further attention.
The tensor-to-scalar ratio is an interesting case. The original matter bounce
predicted the evolution of scalar and tensor modes to be parallel, and therefore, it
predicted a tensor-to-scalar ratio of at least order 1 (the precise value depends on the
convention used to define Ph ). In 2014, the BICEP2 collaboration measured B-mode
polarization in the CMB, which they interpreted as primordial tensor fluctuations.
Their resulting value for the tensor-to-scalar ratio was [67]
r = 0.20+0.07
−0.05 (68% CL) . (2.46)
17
Following the publication of this result, Cai et al. [68] explored the possibility for
several models of bouncing cosmology to reproduce this observational constraint, and
this is the work that we present in Chapter 3.
The validity of the BICEP2 results were disputed after their publication, and
in early 2015, a joint analysis with the Planck collaboration confirmed that most
of the BICEP2 B-mode polarization signal was due to galactic dust [69]. Now, the
consensus is that the tensor-to-scalar ratio is bounded by [69] (see also [64])
r < 0.12 (95% CL) . (2.47)
This result implies that some constraints derived in Chapter 3 are not valid anymore,
but the general discussion of producing r . O(0.1) is still relevant.
Another important aspect that a theory of the very early universe must address
is reheating. To correspond with the content of today’s observable universe, it is
well-known that the universe produced a certain amount of Hydrogen, Helium, and
other light isotopes at early times. This is known as primordial nucleosynthesis.
One constraint from nucleosynthesis is that standard model particles must have
been present early on and at a high enough temperature. The process through
which the hypothetical field (or the fields) responsible for the very early universe
dynamics interacts to produce standard model particles is called reheating. In the
context of inflation, the accelerated expansion of the universe dilutes any form of
matter that would have been already present, and thus, there must be a specific
mechanism through which particles such as photons and electrons are produced (a
more thorough introduction to reheating is given in Chapter 4). In the context of
the matter-Ekpyrotic model, it is the Ekpyrotic phase of contraction which wipes
out any matter content, and thus, requires a reheating mechanism so that enough
radiation is present after the bounce. We explored different reheating mechanisms
in [70] and this is what we present in Chapter 4.
A more theoretical concern about the matter-Ekpyrotic bounce is related to
the stability of the model with respect to quantum corrections. If ordinary field
theories involving polynomial potentials and normal kinetic terms are typically stable
18
against quantum corrections (e.g. the standard model [71]), it is not obvious that
a Lagrangian such as the one introduced in Eq. (2.39) with higher-order derivative
terms and an exponential potential of the form of Eq. (2.42) would be stable against
quantum corrections. This issue certainly deserves closer attention. Let us mention
that if the exponential potential of the Ekpyrotic phase of contraction were unstable
against quantum corrections, it would not necessarily be a fatal problem for the
bounce model. In fact, the Ekpyrotic phase of contraction, which has an equation of
state parameter w 1, is introduced to wash out anisotropies, but this can be done
with any matter having w > 1. Thus, the exponential potential of the Ekpyrotic
model is not crucial to the model. Moreover, the stability of the standard model
Higgs boson seems to indicate that an equation of state w > 1 is naturally reached
at high energy and it has recently been realized that this might play an important
role in bouncing cosmology [72].
To end this section, let us mention additional concerns about the matter-Ekpyrotic
bounce and other bouncing cosmologies. First, a recent article by Battarra et al. [73]
studied the evolution of cosmological perturbations through a ghost condensate non-
singular bounce. The analysis was performed numerically in the harmonic gauge,
and they found no enhancement of the curvature perturbations in the bounce, con-
tradicting the results of Cai et al. [61,63]. This could have profound implications for
the value of the tensor-to-scalar ratio in the matter-Ekpyrotic scenario. Second, re-
cent articles by Gao et al. [74,75] have shown that a particular model of nonsingular
bouncing cosmology produced large amounts of non-gaussianities, beyond current
observational bounds. These non-gaussianities would be produced via the non-linear
evolution of curvature perturbations through the bounce and may be important in
the matter-Ekpyrotic model. We believe that the evolution of curvature perturba-
tions through the bounce and the production of non-gaussianities are linked, and we
will address these issues soon [76].
19
Bibliography
[1] W. H. Kinney, TASI Lectures on Inflation, arXiv:0902.1529 [astro-ph.CO].
[2] P. A. R. Ade et al. [Planck Collaboration], Planck 2015 results. XIII. Cosmological
parameters, arXiv:1502.01589 [astro-ph.CO].
[3] R. H. Brandenberger, Inflationary cosmology: Progress and problems, hep-
ph/9910410.
[4] H. Kodama and M. Sasaki, Cosmological Perturbation Theory, Prog. Theor. Phys.
Suppl. 78, 1 (1984).
[5] V. F. Mukhanov, H. A. Feldman and R. H. Brandenberger, Theory of cosmo-
logical perturbations. Part 1. Classical perturbations. Part 2. Quantum theory of
perturbations. Part 3. Extensions, Phys. Rept. 215, 203 (1992).
[6] R. H. Brandenberger, Lectures on the theory of cosmological perturbations, Lect.
Notes Phys. 646, 127 (2004) [hep-th/0306071].
[7] P. Peter, Cosmological Perturbation Theory, arXiv:1303.2509 [astro-ph.CO].
[8] V. F. Mukhanov, Physical foundations of cosmology, Cambridge University Press,
Cambridge, 2005.
[9] P. Peter and J.-P. Uzan, Primordial Cosmology, Oxford University Press, New
York, 2009.
[10] A. A. Starobinsky, Spectrum of relict gravitational radiation and the early state
of the universe, JETP Lett. 30, 682 (1979) [Pisma Zh. Eksp. Teor. Fiz. 30, 719
(1979)].
[11] W. H. Press, Spontaneous Production of the Zel’dovich Spectrum of Cosmological
Fluctuations, Phys. Scripta 21, 702 (1980).
[12] V. F. Mukhanov and G. V. Chibisov, Quantum Fluctuation and Nonsingular
Universe, JETP Lett. 33, 532 (1981) [Pisma Zh. Eksp. Teor. Fiz. 33, 549 (1981)].
20
21
[13] A. H. Guth and S. Y. Pi, Fluctuations in the New Inflationary Universe, Phys.
Rev. Lett. 49, 1110 (1982).
[14] S. W. Hawking, The Development of Irregularities in a Single Bubble Inflation-
ary Universe, Phys. Lett. B 115, 295 (1982).
[15] A. D. Linde, Scalar Field Fluctuations in Expanding Universe and the New
Inflationary Universe Scenario, Phys. Lett. B 116, 335 (1982).
[16] J. M. Bardeen, P. J. Steinhardt and M. S. Turner, Spontaneous Creation of
Almost Scale - Free Density Perturbations in an Inflationary Universe, Phys.
Rev. D 28, 679 (1983).
[17] D. Wands, Duality invariance of cosmological perturbation spectra, Phys. Rev.
D 60, 023507 (1999) [gr-qc/9809062].
[18] F. Finelli and R. Brandenberger, On the generation of a scale invariant spectrum
of adiabatic fluctuations in cosmological models with a contracting phase, Phys.
Rev. D 65, 103522 (2002) [hep-th/0112249].
[19] S. W. Hawking and G. F. R. Ellis, The Large scale structure of space-time,
Cambridge University Press, Cambridge, 1973.
[20] R. H. Brandenberger, V. F. Mukhanov and A. Sornborger, A Cosmological the-
ory without singularities, Phys. Rev. D 48, 1629 (1993) [gr-qc/9303001].
[21] T. Biswas, A. Mazumdar and W. Siegel, Bouncing universes in string-inspired
gravity, JCAP 0603, 009 (2006) [hep-th/0508194].
[22] A. Kehagias and E. Kiritsis, Mirage cosmology, JHEP 9911, 022 (1999) [hep-
th/9910174].
[23] R. Brandenberger, H. Firouzjahi and O. Saremi, Cosmological Perturbations on
a Bouncing Brane, JCAP 0711, 028 (2007) [arXiv:0707.4181 [hep-th]].
[24] K. Bamba, A. N. Makarenko, A. N. Myagky and S. D. Odintsov, Bounce universe
from string-inspired Gauss-Bonnet gravity, arXiv:1411.3852 [hep-th].
[25] S. D. Odintsov, V. K. Oikonomou and E. N. Saridakis, Superbounce and Loop
Quantum Ekpyrotic Cosmologies from Modified Gravity: F (R), F (G) and F (T )
Theories, arXiv:1501.06591 [gr-qc].
22
[26] J. Haro and J. Amorós, Matter Bounce Scenario in F(T) gravity,

arXiv:1501.06270 [gr-qc].
[27] R. H. Brandenberger, C. Kounnas, H. Partouche, S. P. Patil and N. Toum-
bas, Cosmological Perturbations Across an S-brane, JCAP 1403, 015 (2014)
[arXiv:1312.2524 [hep-th]].
[28] M. Bojowald, Quantum cosmology : A Fundamental Description of the Universe,
Lect. Notes Phys. 835, pp.1 (2011).
[29] A. Ashtekar and P. Singh, Loop Quantum Cosmology: A Status Report, Class.
Quant. Grav. 28, 213001 (2011) [arXiv:1108.0893 [gr-qc]].
[30] E. Wilson-Ewing, The Matter Bounce Scenario in Loop Quantum Cosmology,
JCAP 1303, 026 (2013) [arXiv:1211.6269 [gr-qc]].
[31] E. Wilson-Ewing, Ekpyrotic loop quantum cosmology, JCAP 1308, 015 (2013)
[arXiv:1306.6582 [gr-qc]].
[32] Y. F. Cai and E. Wilson-Ewing, Non-singular bounce scenarios in loop quan-
tum cosmology and the effective field description, JCAP 1403, 026 (2014)
[arXiv:1402.3009 [gr-qc]].
[33] J. Amorós, J. de Haro and S. D. Odintsov, Bouncing loop quantum cosmology
from F (T ) gravity, Phys. Rev. D 87, 104037 (2013) [arXiv:1305.2344 [gr-qc]].
[34] P. Hořava, Quantum Gravity at a Lifshitz Point, Phys. Rev. D 79, 084008 (2009)
[arXiv:0901.3775 [hep-th]].
[35] R. Brandenberger, Matter Bounce in Horava-Lifshitz Cosmology, Phys. Rev. D
80, 043516 (2009) [arXiv:0904.2835 [hep-th]].
[36] B. Feng, X. L. Wang and X. M. Zhang, Dark energy constraints from the cosmic
age and supernova, Phys. Lett. B 607, 35 (2005) [astro-ph/0404224].
[37] B. Feng, M. Li, Y. S. Piao and X. Zhang, Oscillating quintom and the recurrent
universe, Phys. Lett. B 634, 101 (2006) [astro-ph/0407432].
[38] Y. F. Cai, T. Qiu, Y. S. Piao, M. Li and X. Zhang, Bouncing universe with
quintom matter, JHEP 0710, 071 (2007) [arXiv:0704.1090 [gr-qc]].
[39] Y. F. Cai, T. t. Qiu, J. Q. Xia and X. Zhang, A Model Of Inflationary Cosmology
Without Singularity, Phys. Rev. D 79, 021303 (2009) [arXiv:0808.0819 [astro-ph]].
23
[40] Y. F. Cai, T. Qiu, R. Brandenberger, Y. S. Piao and X. Zhang, On Perturbations

of Quintom Bounce, JCAP 0803, 013 (2008) [arXiv:0711.2187 [hep-th]].
[41] Y. F. Cai and X. Zhang, Evolution of Metric Perturbations in Quintom Bounce
model, JCAP 0906, 003 (2009) [arXiv:0808.2551 [astro-ph]].
[42] T. D. Lee and G. C. Wick, Negative Metric and the Unitarity of the S Matrix,
Nucl. Phys. B 9, 209 (1969).
[43] T. D. Lee and G. C. Wick, Finite Theory of Quantum Electrodynamics, Phys.
Rev. D 2, 1033 (1970).
[44] B. Grinstein, D. O’Connell and M. B. Wise, The Lee-Wick standard model,
Phys. Rev. D 77, 025012 (2008) [arXiv:0704.1845 [hep-ph]].
[45] Y. F. Cai, T. t. Qiu, R. Brandenberger and X. m. Zhang, A Nonsingular Cosmol-
ogy with a Scale-Invariant Spectrum of Cosmological Perturbations from Lee-Wick
Theory, Phys. Rev. D 80, 023511 (2009) [arXiv:0810.4677 [hep-th]].
[46] J. M. Cline, S. Jeon and G. D. Moore, The Phantom menaced: Constraints on
low-energy effective ghosts, Phys. Rev. D 70, 043543 (2004) [hep-ph/0311312].
[47] J. Karouby and R. Brandenberger, A Radiation Bounce from the Lee-Wick Con-
struction?, Phys. Rev. D 82, 063532 (2010) [arXiv:1004.4947 [hep-th]].
[48] J. Karouby, T. Qiu and R. Brandenberger, On the Instability of the Lee-Wick
Bounce, Phys. Rev. D 84, 043505 (2011) [arXiv:1104.3193 [hep-th]].
[49] C. Lin, R. H. Brandenberger and L. Perreault Levasseur, A Matter Bounce By
Means of Ghost Condensation, JCAP 1104, 019 (2011) [arXiv:1007.2654 [hep-
th]].
[50] P. Creminelli and L. Senatore, A Smooth bouncing cosmology with scale invariant
spectrum, JCAP 0711, 010 (2007) [hep-th/0702165].
[51] E. I. Buchbinder, J. Khoury and B. A. Ovrut, New Ekpyrotic cosmology, Phys.
Rev. D 76, 123503 (2007) [hep-th/0702154].
[52] T. Qiu, J. Evslin, Y. F. Cai, M. Li and X. Zhang, Bouncing Galileon Cosmolo-
gies, JCAP 1110, 036 (2011) [arXiv:1108.0593 [hep-th]].
[53] D. A. Easson, I. Sawicki and A. Vikman, G-Bounce, JCAP 1111, 021 (2011)
[arXiv:1109.1047 [hep-th]].
24
[54] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, Oscillatory approach to a

singular point in the relativistic cosmology, Adv. Phys. 19, 525 (1970).
[55] D. Battefeld and P. Peter, A Critical Review of Classical Bouncing Cosmologies,
Phys. Rept. 571, 1 (2015) [arXiv:1406.2790 [astro-ph.CO]].
[56] R. H. Brandenberger, The Matter Bounce Alternative to Inflationary Cosmology,
[57] J. Khoury, B. A. Ovrut, P. J. Steinhardt and N. Turok, The Ekpyrotic universe:
Colliding branes and the origin of the hot big bang, Phys. Rev. D 64, 123522
(2001) [hep-th/0103239].
[58] J. Khoury, B. A. Ovrut, N. Seiberg, P. J. Steinhardt and N. Turok, From big
crunch to big bang, Phys. Rev. D 65, 086007 (2002) [hep-th/0108187].
[59] J. L. Lehners, Ekpyrotic and Cyclic Cosmology, Phys. Rept. 465, 223 (2008)
[arXiv:0806.1245 [astro-ph]].
[60] J. K. Erickson, D. H. Wesley, P. J. Steinhardt and N. Turok, Kasner and mix-
master behavior in universes with equation of state w ≥ 1, Phys. Rev. D 69,
063514 (2004) [hep-th/0312009].
[61] Y. F. Cai, D. A. Easson and R. Brandenberger, Towards a Nonsingular Bouncing
Cosmology, JCAP 1208, 020 (2012) [arXiv:1206.2382 [hep-th]].
[62] Y. F. Cai, R. Brandenberger and P. Peter, Anisotropy in a Nonsingular Bounce,
Class. Quant. Grav. 30, 075019 (2013) [arXiv:1301.4703 [gr-qc]].
[63] Y. F. Cai, E. McDonough, F. Duplessis and R. H. Brandenberger, Two Field
Matter Bounce Cosmology, JCAP 1310, 024 (2013) [arXiv:1305.5259 [hep-th]].
[64] P. A. R. Ade et al. [Planck Collaboration], Planck 2015. XX. Constraints on
inflation, arXiv:1502.02114 [astro-ph.CO].
[65] Y. F. Cai and E. Wilson-Ewing, A ΛCDM bounce scenario, JCAP 1503, no.
03, 006 (2015) [arXiv:1412.2914 [gr-qc]].
[66] J. de Haro and Y. F. Cai, An Extended Matter Bounce Scenario: current status
and challenges, arXiv:1502.03230 [gr-qc].
25
[67] P. A. R. Ade et al. [BICEP2 Collaboration], Detection of B-Mode Polarization

at Degree Angular Scales by BICEP2, Phys. Rev. Lett. 112, no. 24, 241101 (2014)
[arXiv:1403.3985 [astro-ph.CO]].
[astro-ph.CO]].
[69] P. A. R. Ade et al. [BICEP2 and Planck Collaborations], Joint Analysis of
BICEP2/KeckArray and P lanck Data, Phys. Rev. Lett. 114, no. 10, 101301
(2015) [arXiv:1502.00612 [astro-ph.CO]].
[70] J. Quintin, Y. F. Cai and R. H. Brandenberger, Matter creation in a nonsingular
bouncing cosmology, Phys. Rev. D 90, no. 6, 063507 (2014) [arXiv:1406.6049 [gr-
qc]].
[71] C. P. Burgess and G. D. Moore, The standard model: A primer, Cambridge,
UK: Cambridge Univ. Pr. (2007) 542 p
[72] R. H. Brandenberger, Y. F. Cai, Y. Wan and X. Zhang, arXiv:1506.06770 [hep-
th].
[73] L. Battarra, M. Koehn, J. L. Lehners and B. A. Ovrut, Cosmological Perturba-
tions Through a Non-Singular Ghost-Condensate/Galileon Bounce, JCAP 1407,
007 (2014) [arXiv:1404.5067 [hep-th]].
[74] X. Gao, M. Lilley and P. Peter, Production of non-gaussianities through a posi-
tive spatial curvature bouncing phase, JCAP 1407, 010 (2014) [arXiv:1403.7958
[gr-qc]].
[75] X. Gao, M. Lilley and P. Peter, Non-Gaussianity excess problem in classical
bouncing cosmologies, Phys. Rev. D 91, no. 2, 023516 (2015) [arXiv:1406.4119
[gr-qc]].
[76] R. H. Brandenberger, Y.-F. Cai, J. Quintin and Z. Sherkatghanad, in prepara-
tion.
Chapter 3
Nonsingular Bouncing Cosmologies in Light of BICEP2
Yi-Fu Cai, Jerome Quintin, Emmanuel N. Saridakis, Edward Wilson-Ewing
Abstract
We confront various nonsingular bouncing cosmologies with the recently released
BICEP2 data and investigate the observational constraints on their parameter space.
In particular, within the context of the effective field approach, we analyze the con-
straints on the matter bounce curvaton scenario with a light scalar field, and the
new matter bounce cosmology model in which the universe successively experiences
a period of matter contraction and an ekpyrotic phase. Additionally, we consider
three nonsingular bouncing cosmologies obtained in the framework of modified grav-
ity theories, namely the Hořava-Lifshitz bounce model, the f (T ) bounce model, and
loop quantum cosmology.
3.1 Introduction
Very recently, the BICEP2 collaboration announced the detection of primor-
dial B-mode polarization in the cosmic microwave background (CMB), claiming an
indirect observation of gravitational waves. This result, if confirmed by other col-
laborations and future observations, will be of major significance for cosmology and
theoretical physics in general. In particular, the BICEP2 team found a tensor-to-
scalar ratio [1]
r = 0.20+0.07
−0.05 , (3.1)
26
at the 1σ confidence level for the ΛCDM scenario. Although there remains the
possibility that the observed B-mode polarization could be partially caused by other
sources [2–4], it is indeed highly probable that the observed B-mode polarization in
the CMB is due at least in part to gravitational waves, remnants of the primordial
universe.
The relic gravitational waves generated in the very early universe is a generic pre-
diction in modern cosmology [5,6]. Inflation is one of several cosmological paradigms
that predicts a roughly scale-invariant spectrum of primordial gravitational waves
[6–8]. The same prediction was also made by string gas cosmology [9–12] and the
matter bounce scenario [13–15]. (Note that the specific predictions of r and the tilt
of the tensor power spectrum can be used in order to differentiate between these
cosmologies.) So far, a lot of the theoretical analyses of the observational data have
been in the context of inflation (see, for instance, [16–34]).
In this present work, we are interested in exploring the consequences of the BI-
CEP2 results in the framework of bouncing cosmological models. In particular, we
desire to study the production of primordial gravitational waves in various bouncing
scenarios, in both the settings of effective field theory and modified gravity. First,
we show that the tensor-to-scalar ratio parameter obtained in a large class of nonsin-
gular bouncing models is predicted to be quite large compared with the observation.
Second, in some explicit models this value can be suppressed due to the nontrivial
physics of the bouncing phase, namely, the matter bounce curvaton [35] and the
new matter bounce cosmology [36–38]. Additionally, for bounce models where the
fluid that dominates the contracting phase has a small sound velocity, primordial
gravitational waves can be generated with very low amplitudes [39]. We show that
the current Planck and BICEP2 data constrain the energy scale at which the bounce
occurs as well as the slope of the Hubble rate during the bouncing phase in these
specific models.
The paper is organized as follows. In Sec. 3.2, we focus on matter bounce cos-
mologies from the effective field theory perspective. In particular, we explore the
matter bounce curvaton scenario [35] and the new matter bounce cosmology [37]. In
Sec. 3.3, we explore another avenue for obtaining nonsingular bouncing cosmologies,
27
that is modifying gravity. We comment on the status of the matter bounce scenario
in Hořava-Lifshitz gravity [40], in f (T ) gravity [41], and in loop quantum cosmol-
ogy [42]. We conclude with a discussion in Sec. 3.4.
3.2 Matter Bounce Cosmology

As an alternative to inflation, the matter bounce cosmology can also give rise
to scale-invariant power spectra for primordial density fluctuations and tensor per-
turbations [13, 14]. In the context of the original matter bounce cosmology, both
the scalar and tensor modes of primordial perturbations grow at the same rate in
the contracting phase before the bounce. As a result, this naturally leads to a large
amplitude of primordial tensor fluctuations [15], greater than the observational up-
per bound. However, an important issue that has to be additionally incorporated in
these calculations is how the perturbations pass through the bouncing phase, which
can drastically decrease the tensor-to-scalar ratio. One example is the matter bounce
in loop quantum cosmology, where the tensor-to-scalar ratio is suppressed by quan-
tum gravity effects during the bounce [43]. Also, for some parameter choices in the
new matter bounce model, r can be suppressed by a small sound speed of the matter
fluid [37].
In this section, we focus on two particularly interesting matter bounce cosmo-
logical models in the effective field theory setting. First, we consider the matter
bounce curvaton model, in which the primordial curvature perturbation can be gen-
erated from the conversion of entropy fluctuations seeded by a second scalar field [35].
Secondly, we investigate the new matter bounce cosmology, where the primordial cur-
vature perturbations can achieve a gravitational amplification during the bounce [36].
3.2.1 The matter bounce curvaton model

The matter bounce curvaton model was originally studied to examine whether
the bouncing solution of the universe is stable against possible entropy fluctua-
tions [35] and particle creation [44]. In a toy model studied in [35], a massless
28
entropy field χ is introduced such that it couples to the bounce field φ via the inter-
action term1 g 2 φ2 χ2 . The entropy field evolves as a tracking solution in the matter
contracting phase and its field fluctuations are nearly scale-invariant provided the
coupling parameter g 2 is sufficiently small. The amplitude of this mode is compa-
rable to the tensor modes and scales as the absolute value of the Hubble parameter
H before the bounce. Afterwards, the universe enters the bouncing phase and the
kinetic term of the entropy field varies rapidly in the vicinity of the bounce. As in
the perturbation equation of motion this term effectively contributes a tachyonic-like
mass, a controlled amplification of the entropy modes can be achieved in the bounc-
ing phase. Since this term does not appear in the equation of motion for tensor
perturbations, the amplitude of primordial gravitational wave is conserved through
this phase. After the bounce, the entropy modes will be transferred into curvature
perturbations, and this increases the amplitude of the power-spectrum of the pri-
mordial density fluctuations. An important consequence of this mechanism is its
suppression of the tensor-to-scalar ratio.
In this subsection, we briefly review the analysis of [35], in light of the BICEP2
results. In the simplest version of the matter bounce curvaton mechanism, there
are only three significant model parameters, namely the coupling parameter g 2 , the
slope parameter of the bouncing phase Υ (which is defined by H ≡ Υt around the
bounce), and the maximal value of the Hubble parameter HB . The value of HB is
associated with the mass of the bounce field m through the following relation
4m
HB ' . (3.2)
3π
The propagation of primordial gravitational waves depends only on the evolution
of the scale factor, and it is possible to calculate the power spectrum for primordial
1
We note that, via this coupling, the massless entropy field χ may obtain a mass
after renormalization. We do not address this point here, but it deserves further
attention in a follow-up study.
29
gravitational waves in this scenario2
2
2Hm
PT = , (3.3)
9π 2 Mp2
from which we see that the amplitude is determined solely by the maximal Hubble
scale Hm . However, the amplitude of the entropy fluctuations is increased during the
bounce. Since tensor perturbations do not couple to scalar perturbations, the entropy
perturbations do not affect the power spectrum of gravitational waves, whereas the
entropy modes are amplified and act as a source for curvature perturbations. This
asymmetry leads to a smaller tensor-to-scalar ratio of
35
r' , (3.4)
F2
where the amplification factor is given by
√
m2

p 3 −1 2 y
F ' exp y(2 + y) + √ sinh , with y ≡ , (3.5)
2 3 Υ
and Υ is the slope parameter of the bouncing phase as defined before Eq. (3.2).
Since the exponent in the above equation is approximately linear in y in the regime
of interest, we see that the tensor-to-scalar ratio can be greatly suppressed for large
values of y, that is for large m or small Υ. Also, we see that r will reach a maximal
value in the massless limit or in the limit where the bounce is instantaneous (i.e.,
Υ → ∞), in which case entropy perturbations are not enhanced.
We recall that, according to the latest observation of the CMB (Planck+WP),
the amplitude of the power spectrum of primordial curvature perturbations is con-
strained to be [45]
ln(1010 As ) = 3.089+0.024
−0.027 (1σ CL) , (3.6)
2
√
Mp ≡ 1/ 8πG is the reduced Planck mass.
30
at the pivot scale k = 0.002 Mpc−1 . Moreover, the recently released BICEP2 data
indicate that [1]
r = 0.20+0.07
−0.05 (1σ CL) . (3.7)
By making use of the above data, we performed a numerical estimate and derived
the constraint on the model parameters Υ and m shown in Fig. 3–1. From the result,
we find that the mass scale m and the slope parameter Υ appearing in the matter
bounce curvaton model have to be in the following ranges
2.5 × 10−4 . m/Mp . 4.5 × 10−4 , (3.8)

7.0 × 10−8 . Υ/Mp2 . 3.5 × 10−7 , (3.9)
respectively. The resulting constraints suggest that if the universe has experienced
a nonsingular matter bounce curvaton, then the energy scale of the bounce should
be of the order of the GUT scale with a smooth and slow bouncing process.
3.2.2 New matter bounce cosmology

In the new matter bounce scenario, as first developed in [36], the universe starts
with a matter-dominated period of contraction and evolves into an ekpyrotic phase
before the bounce. This scenario combines the advantages of the matter bounce
cosmology, which gives rise to scale-invariant primordial power spectra, and the
ekpyrotic universe, which strongly dilutes primordial anisotropies [46]. The model
can be implemented by introducing two scalar fields, as analyzed in the context of
effective field theory [37].
In the effective model of the two field matter bounce [37], one scalar field is intro-
duced to drive the matter-dominated contracting phase and the other is responsible
for the ekpyrotic phase of contraction and the nonsingular bounce. Therefore, similar
to the matter bounce curvaton scenario, there also exist curvature perturbations and
entropy fluctuations during the matter-dominated contracting phase. However, the
main difference between these two models is that, in the present case, the entropy
modes have already been converted into curvature perturbation when the universe
31
4 × 10−7
3 × 10−7
2 × 10−7
Υ/Mp2
10−7
6 × 10−8
3 × 10−4 4 × 10−4
m/Mp
Figure 3–1: Constraints on the mass parameter m and the slope parameter Υ of
the bounce phase from the measurements of Planck and BICEP2 in the matter
bounce curvaton scenario. The blue bands show the 1σ and 2σ confidence intervals
of the tensor-to-scalar ratio and the red bands show the confidence intervals of the
amplitude of the power spectrum of curvature perturbations.
32
enters the ekpyrotic phase before the bounce, while in the matter bounce curvaton
mechanism, this process occurs after the bounce.
In this model, when the universe evolves into the bouncing phase, the kinetic
term of the scalar field that triggers the bounce could vary rapidly which is similar
to the analysis of the matter bounce curvaton mechanism. This process can also ef-
fectively lead to a tachyonic-like mass for curvature perturbations, and therefore, the
corresponding amplitude can be amplified. For the same reason as the matter bounce
curvaton mechanism, this effect only works on the scalar sector. Correspondingly,
the tensor-to-scalar ratio is suppressed when primordial perturbations pass through
the bouncing phase in the new matter bounce cosmology. We would like to point
out that this effect is model dependent, namely, it could be secondary if the ki-
netic term of the background scalar evolves very smoothly compared to the bounce
phase [47, 48]3 .
Following [37], one can write the expression of the power spectrum for primordial
tensor fluctuations as
Fψ2 γψ2 HE2
PT ' , (3.10)
16π 2 (2q − 3)2 Mp2
with
1
γψ ' ,
2(1 − 3q)
√

2 3/2 3
Fψ ' exp 2 ΥtB+ + Υ tB+ , (3.11)
3
up to leading order. In the above expression, HE is the value of the Hubble rate
at the beginning of the ekpyrotic phase and q is an ekpyrotic parameter which is
3
We also want to point out that after the publication of this paper, the paper of
Battarra et al. [49] claimed to find no amplification of the scalar modes compared to
the tensor modes for a similar bouncing model. We plan to clarify the situation in a
future paper [50].
33
much less than unity. Note that we have assumed that the bouncing phase is nearly
symmetric around the bounce point with the values of the scale factor before and
after the bounce being comparable. We denote the time at the end of the bounce
phase by tB+ .
At leading order, the power spectrum for curvature fluctuations is given by
Fζ2 HE2 a2E 2 2

Pζ ' 2 4
γζ m |Uζ |2 , (3.12)
8π Mp
with γζ ' γψ and
HE 27q
Uζ = −(25 + 49q)i − ,
24m 24
√ 2(2 + 3ΥT 2 + Υ2 T 4 ) t3B+

tB+
Fζ ' exp 2 2 + ΥT 2 + √ . (3.13)
T 3 2 + ΥT 2 T3
Similarly to HE , we introduced aE , which is the value of the scale factor at the

beginning of the ekpyrotic phase (in the pre-bounce branch of the universe). Also,
we introduced the mass m of the scalar field responsible for the phase of matter
contraction. We also note the presence of the variable T , which comes into play in
the evolution of the bounce field (see [37] for more details). From eqs. (3.10) and
(3.12), the tensor-to-scalar ratio in this model then takes the form of
PT Fψ2 Mp2
r≡ ' . (3.14)
Pζ (k) 2

2(2q − 3)2 Fζ2 a2E m2 U ζ
Similarly to the previous subsection, we perform a numerical estimate to inves-

tigate the consequences of the observational constraints on the parameter space of
the new matter bounce scenario. Note that, although the model under consideration
involves a series of parameters, there are three main parameters that are most sen-
sitive to observational constraints, i.e., the slope parameter Υ, the Hubble rate at
the beginning of the ekpyrotic phase HE , and the dimensionless duration parameter
tB+ /T of the bouncing phase.
34
7 × 10−7
6 × 10−7
Υ/Mp2
5 × 10−7
4 × 10−7
1.82 1.83 1.84 1.85 1.86 1.87 1.88 1.89 1.90
tB+/T
Figure 3–2: Constraints on the dimensionless duration parameter tB+ /T and the
slope parameter Υ of the bounce phase from the measurements of Planck and BICEP2
in the new matter bounce cosmology. The value of the Hubble parameter at the
beginning of the ekpyrotic phase is fixed to be HE /Mp = 10−7 . As in Fig. 3–1, the
blue and red bands show the confidence intervals of the tensor-to-scalar ratio and of
the amplitude of the power spectrum of curvature perturbations, respectively.
We first look at the correlation between Υ and tB+ /T , with the numerical result
shown in Fig. 3–2. One can read that the slope parameter Υ and the dimensionless
duration parameter tB+ /T are slightly negatively correlated. This implies that one
expects either a slow bounce with a long duration or a fast bounce with a short
duration. However, it is easy to find that the constraint on the dimensionless duration
parameter tB+ /T is very tight with a value slightly less than 2. Therefore, it is
important to examine whether the model predictions accommodate with observations
by fixing the parameter tB+ /T .
35
Then, we analyze the correlation between Υ and HE after setting tB+ /T = 1.86
(found as the best-fit value for the ratio tB+ /T from Fig. 3–2). The allowed parameter
space is depicted by the intersection of the blue and red bands shown in Fig. 3–
3. From the result, we find that the Hubble scale HE and the slope parameter Υ
introduced in the new matter bounce cosmology are constrained to be in the following
ranges
1.9 × 10−8 . HE /Mp . 1.9 × 10−6 , (3.15)

4.9 × 10−7 . Υ/Mp2 . 8.5 × 10−7 , (3.16)
respectively. One can easily see that the constraints on the slope parameter Υ in
the new matter bounce cosmology and in the matter bounce curvaton scenario are
in the same ballpark, i.e., Υ ∼ O(10−7 ) Mp2 . For the new matter bounce cosmology,
if we assume that the bounce occurs at the GUT scale, then the duration of the
bouncing phase is roughly O(104 ) Planck times. We also note that the amplitude
of the Hubble scale HE in the new matter bounce cosmology is much lower than
the GUT scale. This allows for a long enough ekpyrotic contracting phase that can
suppress the unwanted primordial anisotropies as addressed in [46].
In summary, from the analysis of the matter bounce curvaton and the new
matter bounce cosmology scenarios, we can conclude that, in general, a nonsingular
bouncing cosmology has to experience the bouncing phase smoothly for it to agree
with latest observational data. In other words, the Hubble parameter cannot grow
too fast during the bounce phase since the constraints that we find favor a small
value of Υ. Depending on the detailed bounce mechanism, the observed amplitude
of the spectra of the CMB fluctuations may be determined by the bounce scale or the
value of the Hubble parameter at the moment when primordial perturbations were
frozen at super-Hubble scales. For the matter bounce curvaton, the mass scale of
the bounce is of the order of O(10−4 ) Mp which is close to or slightly lower than the
GUT scale (O(1016 ) GeV). On the other hand, for the new matter bounce cosmol-
ogy, due to the introduction of the ekpyrotic phase, it is the Hubble parameter HE
at the onset of the ekpyrotic phase that determines the amplitude of the primordial
36
10−6
9 × 10−7
8 × 10−7
7 × 10−7
Υ/Mp2
6 × 10−7
5 × 10−7
4 × 10−7
10−7 10−6
HE /Mp
Figure 3–3: Constraints on the Hubble parameter at the beginning of the ekpyrotic
phase HE and the slope parameter Υ of the bounce phase from the measurements of
Planck and BICEP2 in the new matter bounce cosmology. The dimensionless bounce
time duration is fixed to be tB+ /T = 1.86. As in Figs. 3–1 and 3–2, the blue and red
bands show the confidence intervals of r and of the amplitude of Pζ , respectively.
37
spectrum of the perturbations, and it must be much lower than the GUT scale in
order to agree with observations. These interesting results encourage further study
of bouncing cosmologies following the effective field approach.
3.3 Implications for modified-gravity bouncing cosmology

In the previous section, we performed numerical computations to constrain two
representative bounce cosmologies that are described by the effective field approach.
It is interesting to extend the analysis to bouncing cosmology models where the
bounce occurs due to modified-gravity theories. In the following, we shall focus on
three specific models. The first one is to obtain the matter bounce solution in the
framework of a non-relativistic modification to Einstein gravity, namely the Hořava-
Lifshitz bounce model [40, 51]. The second is to realize the nonsingular bounce by
virtue of torsion gravity, i.e., the f (T ) bounce model [41]. And the third is a study
of the new matter bounce cosmology in the setting of loop quantum cosmology [42].
3.3.1 Matter bounce in Hořava-Lifshitz gravity

Hořava-Lifshitz gravity is argued to be a potentially UV complete theory for
quantizing the graviton, and it has important implications in the physics of the very
early universe. In particular, a nonsingular bouncing solution can be achieved in this
theory when a non-vanishing spatial curvature term is taken into account [52, 53].
In this case, the higher order spatial derivative terms of the gravity Lagrangian
can effectively contribute a stiff fluid with negative energy, which can trigger the
nonsingular bounce as well as suppressing unwanted primordial anisotropies. Thus,
the bouncing solutions obtained in this picture are marginally stable against the
BKL instability.
As was shown in [40], if the contracting phase is dominated by a pressure-
less matter fluid, Hořava-Lifshitz gravity can provide a realization of the matter
bounce scenario. Moreover, for primordial perturbations in the infrared limit, the
corresponding power spectrum for both the scalar and tensor modes are almost scale-
invariant [54]. However, the paradigm derived in this framework belongs to the tradi-
tional matter bounce cosmology, and so the amplitude of the tensor power spectrum
38
is of the same order as the scalar power spectrum. In this regard, the corresponding
tensor-to-scalar ratio is too large to explain the latest cosmological observations. To
address this issue, one needs to enhance the amplitude of the curvature perturbations
generated in the contracting phase in the infrared limit, for example by applying the
matter bounce curvaton mechanism.
3.3.2 The f (T ) matter bounce cosmology

We briefly describe the realization of the matter bounce in the f (T ) gravitational
modifications of general relativity. The f (T ) gravity theory is a generalization of the
formalism of the teleparallel equivalent of general relativity [55–57], in which one uses
the curvature-less Weitzenböck connection instead of the torsion-less Levi-Civita one,
and thus all of the gravitational information is included in the torsion tensor.
We use the vierbeins eA (xµ ) (Greek indices run over the coordinate space-
time and capital Latin indices run over the tangent space-time) as the dynam-
ical field, related to the metric through gµν (x) = ηAB eA B
µ (x) eν (x), with ηAB =
diag(1, −1, −1, −1). Thus, the torsion tensor is T λµν = eλA (∂µ eA A
ν − ∂ν eµ ), and the
torsion scalar is given by
1 µνλ 1
T = T Tµνλ + T µνλ Tλνµ − Tν νµ T λλµ . (3.17)
4 2
Thus, inspired by the f (R) modifications of Einstein-Hilbert action, one can con-
struct the f (T ) modified gravity by taking the gravity action to be an arbitrary
R
function of the torsion scalar through S = d4 x e[T + f (T )]/16πG. One interest-
ing feature of the f (T ) gravity is that the null energy condition can be effectively
violated, and thus nonsingular bouncing solutions are possible. In particular, it has
been shown that the matter bounce cosmology can be achieved by reconstructing
the form of f (T ) under specific parameterizations of the scale factor [41].
In this model the power spectrum of primordial curvature perturbations is also
scale-invariant if the contracting branch is matter-dominated. Its form is given by
2
Hm
Pζ = , (3.18)
48π 2 Mp2
39
where Hm is the maximal value of the Hubble parameter throughout the whole
evolution [41] (and thus its definition is the same as that introduced in the matter
bounce curvaton scenario).
Now we investigate the evolution of primordial tensor fluctuations in the f (T )
matter bounce. The perturbation equation for the tensor modes can be expressed
as [56]
∇2

12H Ḣf,T T
ḧij + 3H ḣij − 2 hij − ḣij = 0 , (3.19)
a 1 + f,T
where the tensor modes are transverse and traceless. It is interesting to note that
the last term appearing in Eq. (3.19) plays a role of an effective “mass” for the tensor
modes which may affect their amplitudes along the cosmic evolution. However, as
it was pointed out in [41], the effect brought by f,T T is negligible since f (T ) is
approximately a linear function of T in the matter contracting phase. Hence, for
primordial tensor fluctuations at large length scales, although the power spectrum is
also scale-invariant, the amplitude takes of the form of
2
Hm
PT = . (3.20)
2π 2 Mp2
Thus, this result is already ruled out by the present observations unless one intro-
duces some mechanism to magnify the amplitude of scalar-type metric perturbations.
This issue can be resolved by the matter bounce curvaton mechanism. Doing so, the
tensor-to-scalar ratio in our model can be suppressed by the kinetic amplification
factor in the bouncing phase as described in the previous section.
3.3.3 Loop quantum cosmology

The realization of bouncing cosmologies becomes very natural in the frame of
loop quantum cosmology (LQC) since the classical big-bang singularity is generically
replaced by a quantum bounce when the space-time curvature of the universe is of
the order of the Planck scale [58,59]. Several cosmological models have been studied
in the context of LQC, including inflation [60–62], the matter bounce [43], and the
ekpyrotic scenario [63]. Note however that anisotropies are generically expected
40
to become important near the bounce point —with the exception of the ekpyrotic
scenario— and, while the bounce is robust in the presence of anisotropies [64, 65],
the analysis of the cosmological perturbations becomes considerably more complex
when anisotropies are important.
There are two realizations of the matter bounce scenario that have been studied
so far in LQC: the “pure” matter bounce model, where the dynamics are matter-
dominated at all times, including the bounce, and the new matter bounce model
(also called the matter-ekpyrotic bounce) where the space-time is matter-dominated
at the beginning of the contracting phase, while an ekpyrotic scalar field dominates
the dynamics during the end of the contracting era and also the bounce.
In LQC, the dynamics of cosmological perturbations are given by the effective
equation of motion for the Mukhanov-Sasaki variables that include quantum gravity
effects,
zi00

i 00 2 2
vk + cs k − vki = 0 , (3.21)
zi
where k labels the Fourier modes and the index i = {S, T } denotes the scalar and
tensor modes respectively. The detailed forms of the sound speed parameter cs and
the coefficient zi for holonomy-corrected LQC are c2s = 1 − 2ρ/ρc , zT2 = a2 /c2s , and
zS2 = a2 (ρ + P )/H 2 . Here ρc ∼ Mp4 is the critical energy density of LQC where
the bounce occurs. These are the equations of motion that were used to determine
the observational predictions of the pure matter bounce and the matter-ekpyrotic
bounce models in LQC.
In the pure matter bounce model in LQC, tensor perturbations are strongly
suppressed during the bounce due to the quantum gravity modification of zT and
this gives a predicted tensor-to-scalar ratio of r ∼ O(10−3 ), well below the signal
detected by BICEP2 [43]. In addition, the amplitude of the spectrum of scalar
perturbations is of the order of ρc /Mp4 , and therefore in order to match observations,
it is necessary for ρc to be several orders of magnitude below the Planck energy
density. This is problematic as heuristic arguments relating LQC and the full theory
41
of loop quantum gravity indicate that ρc is expected to be at most one or two orders
of magnitude below ρPl .
This last problem is avoided in the new matter bounce model for the follow-
ing reason: when the universe evolves into the ekpyrotic phase, all the perturbation
modes at super-Hubble scales freeze and thus the amplitude of the perturbations
are entirely determined by the value of the Hubble parameter at the beginning of
the ekpyrotic phase, HE [42]. Because of this, the observed amplitude of the scalar
perturbations determines HE , not ρc . In addition, the ekpyrotic phase also dilutes
the anisotropies before the bounce occurs and hence the BKL instability is avoided
in this model. In order to determine how the recent results of the BICEP2 collab-
oration constrain the matter-ekpyrotic bounce in LQC, it is necessary to determine
the amplitude of the primordial tensor fluctuations in this scenario.
The dynamics of scalar perturbations in the LQC matter-ekpyrotic bounce
model have been studied in detail in [42], and this analysis is easy to extend to
tensor perturbations as their evolution is given by a very similar differential equa-
tion as seen in Eq. (3.21). Due to the fact that their equations of motion are very
similar at times well before the bounce, the amplitude of the spectra of the scalar
and tensor modes are of the same order, and it is easy to check that if the ekpyrotic
scalar field dominates the dynamics during the bounce,both the scalar and tensor
modes evolve trivially through the bounce (note that this is very different to what
happens if the matter field dominates the dynamics during the bounce). The result
is that, as in other matter-ekpyrotic bounce models without entropy perturbations,
the resulting amplitude of the tensor perturbations is significantly larger than for the
scalar perturbations and therefore this particular model is ruled out by observations.
However, if there is more than one matter field then entropy perturbations may
become important, and they have been neglected in the above analysis. As explained
in Sec. 3.2, entropy perturbations can significantly increase the amplitude of scalar
perturbations, while not affecting the dynamics of tensor perturbations in any way,
thus decreasing the tensor-to-scalar ratio. Therefore, for the matter-ekpyrotic bounce
model to be viable in LQC, it will be necessary to include entropy perturbations in
42
some manner, perhaps as is done in the new matter bounce model presented in Sec.
3.2.2.
Finally, it is possible (at least for the flat FLRW space-time) to express LQC as
a teleparallel theory, which leads to slightly different equations of motion for cosmo-
logical perturbations [66]. In this setting, as there exist solutions with a large range
of tensor-to-scalar ratios, r ∈ [0.1243, 13.4375] [67], it is possible to obtain a value of
r that is compatible with the results of the BICEP2 collaboration.
3.4 Conclusion
In this work, we confronted various bouncing cosmologies with the recently
released BICEP2 data. In particular, we analyzed two scenarios in the effective
field theory framework, namely the matter bounce curvaton scenario and new mat-
ter bounce cosmology, and three modified gravity theories, namely Hořava-Lifshitz
gravity, the f (T ) theories, and loop quantum cosmology. In all of these models, we
showed their capability of generating primordial gravitational waves.
Since matter bounce models typically produce a large amount of primordial ten-
sor fluctuations, specific mechanisms for their suppression are needed. In the matter
bounce curvaton scenario, introducing an extra scalar coupled to the bouncing field
induces a controllable amplification of the entropy modes during the bouncing phase,
and since these modes will be transferred into curvature perturbations the resulting
tensor-to-scalar ratio is suppressed to a value in agreement with the observations of
the BICEP2 collaboration.
Another possibility, called the new matter bounce cosmology, is to have two
scalar fields, one driving the matter contracting phase and the other driving the
ekpyrotic contraction and the nonsingular bounce. Thus, the entropy modes are
converted into curvature perturbations when the universe enters the ekpyrotic phase
before the bounce, and the resulting tensor-to-scalar ratio is again suppressed to
observed values.
Furthermore, in both of these models we used the BICEP2 and the Planck results
in order to constrain the free parameters in these models, namely the energy scale of
the bounce, the slope of the Hubble rate during the bouncing phase, or the Hubble
43
rate at the beginning of the ekpyrotic-dominated phase for the new matter bounce
cosmology.
Finally, we considered bouncing cosmologies in the framework of modified grav-
ity. In particular, in both the Hořava-Lifshitz bounce model and the f (T ) gravity
bounce, we have argued that the presence of a curvaton field may suppress the tensor-
to-scalar ratio to its observed values. We leave the detailed analysis of this topic for
a follow-up study.
In loop quantum cosmology, two realizations of the matter bounce have been
studied. In the simplest matter bounce model where there is only one matter field, the
amplitude of the tensor perturbations is significantly diminished during the bounce
due to quantum gravity effects; this process predicts a very small value of r ∼
O(10−3 ), well below the value observed by BICEP2. The other model that has
been studied is the new matter bounce scenario, which in the absence of entropy
perturbations predicts a large amplitude for the tensor perturbations (in this case
quantum gravity effects do not modify the spectrum during the bounce). Therefore,
for the new matter bounce scenario in LQC to be viable, it is also necessary to include
entropy perturbations in order to lower the value of r to a value in agreement with the
results of BICEP2. Also, as can be seen here, the dominant field during the bounce
significantly affects how the value of r changes during the bounce and therefore it
seems likely that by carefully choosing this field, it may be possible to obtain a
tensor-to-scalar ratio in agreement with observations. We leave this possibility for
future work.
In summary, the predictions of the matter bounce cosmologies where entropy
perturbations significantly increase the amplitude of scalar perturbations remain
consistent with observations, and thus these models are good alternatives to inflation.
44
ACKNOWLEDGEMENTS
We are indebted to Robert Brandenberger and Jean-Luc Lehners for valuable

comments. We also thank Taotao Qiu for helpful discussions in the initial stages of
the project. The work of YFC and JQ is supported in part by NSERC and by funds
from the Canada Research Chair program. The research of ENS is implemented
within the framework of the Action “Supporting Postdoctoral Researchers” of the
Operational Program “Education and Lifelong Learning” (Actions Beneficiary: Gen-
eral Secretariat for Research and Technology), and is co-financed by the European
Social Fund (ESF) and the Greek State. The work of EWE is supported in part
by a grant from the John Templeton Foundation. The opinions expressed in this
publication are those of the authors and do not necessarily reflect the views of the
John Templeton Foundation.
45
Bibliography
[1] P. A. R. Ade et al. [BICEP2 Collaboration], Detection of B-Mode Polarization
at Degree Angular Scales by BICEP2, Phys. Rev. Lett. 112, no. 24, 241101
(2014) [arXiv:1403.3985 [astro-ph.CO]].
[2] J. Lizarraga, J. Urrestilla, D. Daverio, M. Hindmarsh, M. Kunz and
A. R. Liddle, Can topological defects mimic the BICEP2 B-mode signal?,
Phys. Rev. Lett. 112, 171301 (2014) [arXiv:1403.4924 [astro-ph.CO]].
[3] A. Moss and L. Pogosian, Did BICEP2 see vector modes? First B-mode
constraints on cosmic defects, Phys. Rev. Lett. 112, 171302 (2014)
[4] C. Bonvin, R. Durrer and R. Maartens, Can primordial magnetic fields be the
origin of the BICEP2 data?, Phys. Rev. Lett. 112, no. 19, 191303 (2014)
[5] L. P. Grishchuk, Amplification of gravitational waves in an istropic universe,
Sov. Phys. JETP 40, 409 (1975) [Zh. Eksp. Teor. Fiz. 67, 825 (1974)].
[6] A. A. Starobinsky, Spectrum of relict gravitational radiation and the early state
of the universe, JETP Lett. 30, 682 (1979) [Pisma Zh. Eksp. Teor. Fiz. 30,
719 (1979)].
[7] V. A. Rubakov, M. V. Sazhin and A. V. Veryaskin, Graviton Creation in the
Inflationary Universe and the Grand Unification Scale, Phys. Lett. B 115, 189
(1982).
[8] A. A. Starobinsky, Cosmic Background Anisotropy Induced by Isotropic
Flat-Spectrum Gravitational-Wave Perturbations, Sov. Astron. Lett. 11, 133
(1985).
[9] R. H. Brandenberger and C. Vafa, Superstrings in the Early Universe, Nucl.
Phys. B 316, 391 (1989).
46
47
[10] R. H. Brandenberger, A. Nayeri, S. P. Patil and C. Vafa, Tensor Modes from a

Primordial Hagedorn Phase of String Cosmology, Phys. Rev. Lett. 98, 231302
(2007) [hep-th/0604126].
[11] R. H. Brandenberger, A. Nayeri and S. P. Patil, Closed String
Thermodynamics and a Blue Tensor Spectrum, Phys. Rev. D 90, no. 6, 067301
(2014) [arXiv:1403.4927 [astro-ph.CO]].
[12] T. Biswas, T. Koivisto and A. Mazumdar, Atick-Witten Hagedorn Conjecture,
near scale-invariant matter and blue-tilted gravity power spectrum, JHEP
1408, 116 (2014) [arXiv:1403.7163 [hep-th]].
D 60, 023507 (1999) [gr-qc/9809062].
[14] F. Finelli and R. Brandenberger, On the generation of a scale invariant
spectrum of adiabatic fluctuations in cosmological models with a contracting
phase, Phys. Rev. D 65, 103522 (2002) [hep-th/0112249].
[15] Y. F. Cai, T. t. Qiu, R. Brandenberger and X. m. Zhang, A Nonsingular
Cosmology with a Scale-Invariant Spectrum of Cosmological Perturbations from
Lee-Wick Theory, Phys. Rev. D 80, 023511 (2009) [arXiv:0810.4677 [hep-th]].
[16] A. Kehagias and A. Riotto, Remarks about the Tensor Mode Detection by the
BICEP2 Collaboration and the Super-Planckian Excursions of the Inflaton
Field, Phys. Rev. D 89, no. 10, 101301 (2014) [arXiv:1403.4811 [astro-ph.CO]].
[17] Y. Z. Ma and Y. Wang, Reconstructing the Local Potential of Inflation with
BICEP2 data, JCAP 1409, no. 09, 041 (2014) [arXiv:1403.4585 [astro-ph.CO]].
[18] K. Harigaya and T. T. Yanagida, Discovery of Large Scale Tensor Mode and
Chaotic Inflation in Supergravity, Phys. Lett. B 734, 13 (2014)
[arXiv:1403.4729 [hep-ph]].
[19] J. O. Gong, Blue running of the primordial tensor spectrum, JCAP 1407, 022
(2014) [arXiv:1403.5163 [astro-ph.CO]].
[20] V. Miranda, W. Hu and P. Adshead, Steps to Reconcile Inflationary Tensor
and Scalar Spectra, Phys. Rev. D 89, no. 10, 101302 (2014) [arXiv:1403.5231
[astro-ph.CO]].
[21] M. P. Hertzberg, Inflation, Symmetry, and B-Modes, arXiv:1403.5253 [hep-th].
48
[22] D. H. Lyth, BICEP2, the curvature perturbation and supersymmetry, JCAP

1411, no. 11, 003 (2014) [arXiv:1403.7323 [hep-ph]].
[23] J. Q. Xia, Y. F. Cai, H. Li and X. Zhang, Evidence for bouncing evolution
before inflation after BICEP2, Phys. Rev. Lett. 112, 251301 (2014)
[24] D. K. Hazra, A. Shafieloo, G. F. Smoot and A. A. Starobinsky, Inflation with
Whip-Shaped Suppressed Scalar Power Spectra, Phys. Rev. Lett. 113, no. 7,
071301 (2014) [arXiv:1404.0360 [astro-ph.CO]].
[25] Y. F. Cai, J. O. Gong and S. Pi, Inflation beyond T-models and primordial
B-modes, Phys. Lett. B 738, 20 (2014) [arXiv:1404.2560 [hep-th]].
[26] M. W. Hossain, R. Myrzakulov, M. Sami and E. N. Saridakis, Class of
quintessential inflation models with parameter space consistent with BICEP2,
Phys. Rev. D 89, no. 12, 123513 (2014) [arXiv:1404.1445 [gr-qc]].
[27] B. Hu, J. W. Hu, Z. K. Guo and R. G. Cai, Reconstruction of the primordial
power spectra with Planck and BICEP2 data, Phys. Rev. D 90, no. 2, 023544
(2014) [arXiv:1404.3690 [astro-ph.CO]].
[28] W. Zhao, C. Cheng and Q. G. Huang, Hint of relic gravitational waves in the
Planck and WMAP data, arXiv:1403.3919 [astro-ph.CO].
[29] J. F. Zhang, Y. H. Li and X. Zhang, Sterile neutrinos help reconcile the
observational results of primordial gravitational waves from Planck and
BICEP2, Phys. Lett. B 740, 359 (2015) [arXiv:1403.7028 [astro-ph.CO]].
[30] P. Di Bari, S. F. King, C. Luhn, A. Merle and A. Schmidt-May, Radiative
Inflation and Dark Energy RIDEs Again after BICEP2, JCAP 1408, 040
(2014) [arXiv:1404.0009 [hep-ph]].
[31] H. Li, J. Q. Xia and X. Zhang, Global fitting analysis on cosmological models
after BICEP2, arXiv:1404.0238 [astro-ph.CO].
[32] Y. C. Chung and C. Lin, Topological Inflation with Large Tensor-to-scalar
Ratio, JCAP 1407, 020 (2014) [arXiv:1404.1680 [astro-ph.CO]].
[33] Y. F. Cai and Y. Wang, Testing quantum gravity effects with latest CMB
observations, Phys. Lett. B 735, 108 (2014) [arXiv:1404.6672 [astro-ph.CO]].
49
[34] M. W. Hossain, R. Myrzakulov, M. Sami and E. N. Saridakis, Evading Lyth

bound in models of quintessential inflation, Phys. Lett. B 737, 191 (2014)
[arXiv:1405.7491 [gr-qc]].
[35] Y. F. Cai, R. Brandenberger and X. Zhang, The Matter Bounce Curvaton
Scenario, JCAP 1103, 003 (2011) [arXiv:1101.0822 [hep-th]].
[36] Y. F. Cai, D. A. Easson and R. Brandenberger, Towards a Nonsingular
Bouncing Cosmology, JCAP 1208, 020 (2012) [arXiv:1206.2382 [hep-th]].
[37] Y. F. Cai, E. McDonough, F. Duplessis and R. H. Brandenberger, Two Field
[38] Y. F. Cai, Exploring Bouncing Cosmologies with Cosmological Surveys, Sci.
China Phys. Mech. Astron. 57, 1414 (2014) [arXiv:1405.1369 [hep-th]].
[39] D. Bessada, N. Pinto-Neto, B. B. Siffert and O. D. Miranda, Stochastic
background of relic gravitons in a bouncing quantum cosmological model, JCAP
1211, 054 (2012) [arXiv:1207.5863 [gr-qc]].
80, 043516 (2009) [arXiv:0904.2835 [hep-th]].
[41] Y. F. Cai, S. H. Chen, J. B. Dent, S. Dutta and E. N. Saridakis, Matter
Bounce Cosmology with the f(T) Gravity, Class. Quant. Grav. 28, 215011
(2011) [arXiv:1104.4349 [astro-ph.CO]].
[42] Y. F. Cai and E. Wilson-Ewing, Non-singular bounce scenarios in loop
quantum cosmology and the effective field description, JCAP 1403, 026 (2014)
[arXiv:1402.3009 [gr-qc]].
JCAP 1303, 026 (2013) [arXiv:1211.6269 [gr-qc]].
[44] Y. F. Cai, R. Brandenberger and X. Zhang, Preheating a bouncing universe,
Phys. Lett. B 703, 25 (2011) [arXiv:1105.4286 [hep-th]].
[45] P. A. R. Ade et al. [Planck Collaboration], Planck 2013 results. XVI.
Cosmological parameters, Astron. Astrophys. 571, A16 (2014)
50
[46] Y. F. Cai, R. Brandenberger and P. Peter, Anisotropy in a Nonsingular

Bounce, Class. Quant. Grav. 30, 075019 (2013) [arXiv:1301.4703 [gr-qc]].
[47] M. Osipov and V. Rubakov, Galileon bounce after ekpyrotic contraction, JCAP
1311, 031 (2013) [arXiv:1303.1221 [hep-th]].
[48] M. Koehn, J. L. Lehners and B. A. Ovrut, Cosmological super-bounce, Phys.
Rev. D 90, no. 2, 025005 (2014) [arXiv:1310.7577 [hep-th]].
[49] L. Battarra, M. Koehn, J. L. Lehners and B. A. Ovrut, Cosmological
Perturbations Through a Non-Singular Ghost-Condensate/Galileon Bounce,
JCAP 1407, 007 (2014) [arXiv:1404.5067 [hep-th]].
[50] R. H. Brandenberger, Y.-F. Cai, J. Quintin and Z. Sherkatghanad, in
preparation.
[51] Y. F. Cai and E. N. Saridakis, Non-singular cosmology in a model of
non-relativistic gravity, JCAP 0910, 020 (2009) [arXiv:0906.1789 [hep-th]].
[52] G. Calcagni, Cosmology of the Lifshitz universe, JHEP 0909, 112 (2009)
[arXiv:0904.0829 [hep-th]].
[53] E. Kiritsis and G. Kofinas, Horava-Lifshitz Cosmology, Nucl. Phys. B 821, 467
(2009) [arXiv:0904.1334 [hep-th]].
[54] Y. F. Cai and X. Zhang, Primordial perturbation with a modified dispersion
relation, Phys. Rev. D 80, 043520 (2009) [arXiv:0906.3341 [astro-ph.CO]].
[55] E. V. Linder, Einstein’s Other Gravity and the Acceleration of the Universe,
Phys. Rev. D 81, 127301 (2010) [Erratum-ibid. D 82, 109902 (2010)]
[56] S. H. Chen, J. B. Dent, S. Dutta and E. N. Saridakis, Cosmological
perturbations in f(T) gravity, Phys. Rev. D 83, 023508 (2011) [arXiv:1008.1250
[astro-ph.CO]].
[57] J. B. Dent, S. Dutta and E. N. Saridakis, f(T) gravity mimicking dynamical
dark energy. Background and perturbation analysis, JCAP 1101, 009 (2011)
[58] A. Ashtekar, T. Pawlowski and P. Singh, Quantum Nature of the Big Bang:
Improved dynamics, Phys. Rev. D 74, 084003 (2006) [gr-qc/0607039].
51
[59] P. Singh, Are loop quantum cosmos never singular?, Class. Quant. Grav. 26,
125005 (2009) [arXiv:0901.2750 [gr-qc]].
[60] A. Ashtekar and D. Sloan, Loop quantum cosmology and slow roll inflation,
Phys. Lett. B 694, 108 (2010) [arXiv:0912.4093 [gr-qc]].
[61] L. Linsefors, T. Cailleteau, A. Barrau and J. Grain, Primordial tensor power
spectrum in holonomy corrected Ω loop quantum cosmology, Phys. Rev. D 87,
no. 10, 107503 (2013) [arXiv:1212.2852 [gr-qc]].
[62] I. Agullo, A. Ashtekar and W. Nelson, The pre-inflationary dynamics of loop
quantum cosmology: Confronting quantum gravity with observations, Class.
Quant. Grav. 30, 085014 (2013) [arXiv:1302.0254 [gr-qc]].
[63] E. Wilson-Ewing, Ekpyrotic loop quantum cosmology, JCAP 1308, 015 (2013)
[arXiv:1306.6582 [gr-qc]].
[64] A. Ashtekar and E. Wilson-Ewing, Loop quantum cosmology of Bianchi I
models, Phys. Rev. D 79, 083535 (2009) [arXiv:0903.3397 [gr-qc]].
[65] B. Gupt and P. Singh, Quantum gravitational Kasner transitions in Bianchi-I
spacetime, Phys. Rev. D 86, 024034 (2012) [arXiv:1205.6763 [gr-qc]].
[66] J. Haro, Cosmological perturbations in teleparallel Loop Quantum Cosmology,
JCAP 1311, 068 (2013) [Erratum-ibid. 1405, E01 (2014)] [arXiv:1309.0352
[gr-qc]].
[67] J. de Haro and J. Amoros, Viability of the matter bounce scenario in Loop
Quantum Cosmology from BICEP2 last data, JCAP 1408, 025 (2014)
[arXiv:1403.6396 [gr-qc]].
Chapter 4
Matter Creation in a Nonsingular Bouncing Cosmology
Jerome Quintin, Yi-Fu Cai, Robert H. Brandenberger
Abstract
We examine reheating in the two-field matter bounce cosmology. In this model,
the Universe evolves from a matter-dominated phase of contraction to an Ekpy-
rotic phase of contraction before the nonsingular bounce. The Ekpyrotic phase frees
the model from unwanted anisotropies, but leaves the Universe cold and empty of
particles after the bounce. For this reason, we explore two particle production mech-
anisms which take place during the course of the cosmological evolution: Parker
particle production where the matter field couples only to gravity and particle cre-
ation via interactions between the matter field and the bounce field. Although we
show that both mechanisms can produce particles in this model, we find that Parker
particle production is sufficient to reheat the Universe to high temperatures. Thus
there is a priori no need to add an interaction term to the Lagrangian of the model.
Still, particle creation via interactions can contribute to the formation of matter and
radiation, but only if the coupling between the fields is tuned to be large.
4.1 Introduction
Reheating (see e.g. [1] for a recent review) is an integral part of inflationary
cosmology. Without a mechanism to convert the energy residing in the inflaton
field at the end of inflation into regular matter, inflation would produce a universe
devoid of any regular matter and radiation, obviously a cosmological catastrophe.
52
In a similar way, an early universe cosmology alternative to inflation must contain a
mechanism which leads to regular matter and radiation at late times.
The matter production mechanisms will differ from model to model. In infla-
tionary cosmology, it is a weak coupling between the inflaton field and the fields
representing regular matter which leads to a parametric resonance (“preheating”) in-
stability in the equation of motion of these matter fields during the time after the end
of inflation when the inflaton is oscillating about its ground state [2–5]. In “string
gas cosmology” [6], it is the annihilation of string winding modes into string loops
which automatically leads to the generation of matter and radiation at the end of
the initial string phase (the “Hagedorn” phase — see e.g. [7] for recent reviews on
string gas cosmology). In the “emergent Galileon cosmology” [8] it is a “defrosting”
transition analogous to the resonant instability at the end of inflation which leads
to the production of regular matter [9]. This mechanism also requires a coupling
between regular matter fields and the Galileon condensate.
A nonsingular bouncing cosmology with a matter-dominated initial phase of
contraction (during which scales of cosmological interest exit the Hubble radius) is
another alternative to cosmological inflation for producing the spectrum of cosmo-
logical fluctuations observed today [10, 11]. The nonsingularity of the bounce can
be obtained by introducing new physics in the matter sector such as a quintom
field [12] or a ghost condensate [13], or by modifying gravity at high scales, e.g. in
the context of string theory [14], Horava-Lifshitz gravity [15], nonlocal gravity [16],
or loop quantum cosmology [17]. If there is no additional phase of contraction dur-
ing which the relative contribution of radiation to the total energy density decreases,
then there is no need for a matter generation mechanism during and after the bounce
since the original matter and radiation content of the Universe at early times during
the contracting phase is preserved. However, it is precisely such bouncing models
without a dilution mechanism during contraction which suffer from the “anisotropy
problem” [“Belinsky-Khalatnikov-Lifshitz (BKL) instability” [18]; the energy density
in the anisotropies grows faster than the energy density in matter and radiation and
destroys the homogeneous bounce]. To solve this anisotropy problem, a model with
a phase of Ekpyrotic contraction was introduced [19, 20] and it was shown explicitly
53
to be stable against anisotropies in [21] (see [22] for a review). However, during
the period of Ekpyrotic contraction the regular matter and radiation become irrel-
evant analogously to what happens to preinflationary matter and radiation during
inflationary expansion, and hence a mechanism is required to recreate matter and
radiation during and after the bounce. This is the topic we address in the present
study.
In this paper, we study gravitational matter field particle creation in the two-field
Ekpyrotic matter bounce model of [20]. This mechanism is called “Parker particle
production” [23–25] (see also [26–30] for early developments; see [31–41] for studies
in the context of inflation; and see [42–44] for recent reviews). Due to the nontrivial
dynamics of the matter mode functions in the evolving background space-time, we
find that particles of matter fields can be efficiently generated. We evaluate the
particle number at the moment of matter-Ekpyrotic equality, and at the beginning
and end of the bouncing phase. Our results explicitly show that by the time of the end
of the bouncing phase there is sufficient gravitational particle production, and that
there is hence no need to introduce any preheating phase. This is very different from
the situation in inflationary cosmology where gravitational particle production is
negligible compared to the energy transfer via the preheating instability. We further
show that even if we introduce a small coupling between the scalar field leading to
Ekpyrotic contraction and the regular matter fields, then parametric resonance is in
fact negligible.
The paper is organized as follows. In Sec. II, we briefly review the model of
two-field matter bounce cosmology proposed in [20]. In Sec. III, we study gravita-
tional particle creation of a regular scalar matter field ψ by tracking the background
evolution. In particular, we present the quantization process of the ψ particles in the
matter contraction, Ekpyrotic contraction, and the bouncing phases, respectively.
Afterwards, we compute the Bogoliubov coefficients of the ψ particles by following
the background evolution, and we determine the magnitude of the energy density in
the produced particles. To double check the validity of the analyses of Sec. III, we
analyze in Sec. IV the same setup by computing the backreaction of the ψ fluctua-
tions on the energy-momentum tensor. Then in Sec. V, we turn on an interaction
54
term coupling the matter field to the field yielding the Ekpyrotic contraction, and
we investigate particle production through direct interactions. We conclude with a
discussion in Sec. VI. Throughout the paper we take the sign of the metric to be
√
(+, −, −, −) and define the reduced Planck mass by Mp = 1/ 8πG.
4.2 The Two-Field Matter Bounce Model

The two-field matter bounce [20] is a simple toy model which yields regular
matter and can give both an Ekpyrotic contraction and a nonsingular bounce. In
this section, we briefly review the dynamics of this nonsingular bounce model.
Following [20], we consider the Lagrangian of the model to be
L = K(φ, X) + G(X)φ + P (ψ, Y ) , (4.1)
where φ is the scalar field responsible for both the Ekpyrotic phase of contraction and
the bounce, and ψ is a second scalar field representing the regular matter component
responsible for the initial matter contraction phase.
For the scalar field φ, we introduce a K-essence-type term K which is a function
of φ and its kinetic term [45]
1 µν
X ≡ g ∂µ φ∂ν φ , (4.2)
2
and a Horndeski-type [46] term Gφ with
≡ g µν ∇µ ∇ν (4.3)
being the standard d’Alembertian operator. When the Universe is far away from the
bounce, the Lagrangian of φ takes the canonical form which is the sum of a regular
kinetic term X and a potential V (φ). More specifically, we choose
K(φ, X) = Mp2 [1 − g(φ)]X + βX 2 − V (φ) (4.4)
and
G(X) = γX , (4.5)
55
with β and γ being two positive-definite constants, and
2g0
g(φ) = √ √ , (4.6)
e− 2/pφ
+ ebg 2/pφ
where g0 > 1, p, and bg are further positive constants. If the potential is chosen
to be a negatively valued exponential function, one can get a phase of Ekpyrotic
contraction [47] which will dilute the unwanted anisotropies [48]. The choice made
in [20] for the potential was
2V0
V (φ) = √ √ , (4.7)
e− 2/qφ
+ ebV 2/qφ
where V0 , q, and bV are more positive constants. As a result, this type of model can
avoid the BKL instability. The initial conditions for φ will be taken to be a large
negative value with small time derivative. In this case, φ will contribute negligibly
to the energy and momentum of the background. As φ increases, it will eventually
reach a value for which g(φ) > 1 and hence a ghost condensate [49] will form for
a short while. This gives rise to a nonsingular bounce via the violation of the null
energy condition.
Now we consider the dynamics of the massive scalar field ψ which represents the
regular matter component. We take the Lagrangian to be of canonical form as
1
P (ψ, Y ) = Y − m2 ψ 2 , (4.8)
2
with Y being defined as the kinetic term
1 µν
Y ≡ g ∂µ ψ∂ν ψ (4.9)
2
and m being the mass of the scalar. When ψ oscillates around its vacuum hψi = 0,
its contribution to the energy-momentum tensor is that of a pressureless fluid. If it
dominates the total energy-momentum tensor, then it leads to a matter-dominated
evolution.
Let us take a universe filled with the above two fields and consider a contracting
universe with initial conditions in which the contribution of the Ekpyrotic field to
56
the energy-momentum tensor is negligible since φ is initialized to have large nega-
tive value and small velocity. Hence, the Universe will undergo a phase of matter-
dominated contraction with an effective background equation of state w = 0. Since
the energy density in the Ekpyrotic field increases faster than that in ψ, the Universe
will at some point enter a period of Ekpyrotic contraction driven by the scalar field
φ, with w evolving to w 1. Afterwards, the Universe experiences a nonsingular
bounce due to the emergence of the ghost condensate phase. The ghost conden-
sate phase ends naturally once φ exceeds a critical positive value. At that point
traditional cosmological thermal expansion takes over. The space-time diagram of
our nonsingular bounce cosmology is sketched in Fig. 4–1, in which the vertical axis
denotes time, and the horizontal axis denotes a comoving space coordinate. The
bounce point is taken to be t = 0. The wavelength of one mode is depicted (the
vertical line), and the Hubble radius is shown. It decreases in comoving coordinates
during the contraction phase and then increases again in the expanding phase.
Since in the two-field matter bounce model we use the Ekpyrotic phase of con-
traction to dilute the undesired anisotropies, this phase also washes out the primor-
dial matter particles. Therefore, one must be concerned that the Universe might
be empty of regular matter and radiation after the bounce. As discussed in the
Introduction, a similar issue arises within the framework of inflationary cosmology.
In the case of inflation, in order to have efficient production of regular matter after
the phase of exponential expansion, one usually introduces an explicit coupling term
between the field φ which yields inflation and regular matter fields. In this case,
already at the level of quantum field perturbation theory one expects the production
of matter particles [50, 51] after inflation has ended, i.e. in a separate phase called
the “reheating period.” However, it was observed in [2] (see also [52]) that the initial
transfer of energy to matter fields is dominated by a parametric resonance instability.
This “preheating period” was further analyzed in [3–5].
The question now arises as to whether it is necessary to introduce an analo-
gous reheating phase in a nonsingular matter bounce scenario in order to explain the
origin of regular matter and radiation. It is obviously possible to introduce such a
57
Figure 4–1: Space-time sketch in the two-field matter bounce scenario. The comoving
length and conformal time are labeled by x and η, respectively. The solid blue curve
shows H−1 , the comoving Hubble radius. The different phase transition times are
labeled on the vertical axis. The shaded areas show the regions of integration for
the respective phases (the Ekpyrotic and bounce phases are shown). The comoving
wavelengths of the fluctuating modes that we integrate originate in these shaded
areas. The green line shows the comoving wavelength of such a fluctuating mode
that forms in the Ekpyrotic phase of contraction. We see that this mode reenters
the Hubble horizon at later times, after ηB+ .
58
mechanism, and this was done within the framework of nonsingular bouncing cos-
mologies in order to enhance the amplitude of the scalar cosmological perturbations
relative to the tensor modes [53]. However, since the scalar modes in fact experience
a substantial growth in amplitude during the bouncing phase (greater than that of
the tensor modes) [20] (see also [54] for earlier work on fluctuations in a nonsingu-
lar bouncing cosmology), it is not necessary to invoke bounce preheating to boost
the scalar fluctuations, and it would be more elegant not to do so since the bounce
preheating scenario requires additional couplings (like inflationary reheating does).
In this present work, we analyze matter production in the two-field matter
bounce model and show that there is no need to introduce preheating. Gravita-
tional particle production is in fact sufficiently effective. Specifically, we consider
the production of ψ particles in our cosmological background. We are interested in
particles with momenta which are large (compared to the Hubble rate in the post-
bounce universe) rather than the large wavelength fluctuations which contribute to
cosmological perturbations.
4.3 Gravitational particle production throughout the cosmic evolution

In an expanding universe the concept of the vacuum state of a quantum field
is not well defined. In Minkowski space-time quantum field theory we can expand
the field in creation and annihilation operators which are the operator coefficients
of the positive and negative frequency modes, respectively. The vacuum state is
then the state which is annihilated by all of the annihilation operators. In curved
space-time there is no global definition of positive and negative frequency modes.
A mode which to an initial observer looks pure positive frequency will look like a
combination of positive and negative frequency modes for a later time observer. The
late time observer thus sees the state as containing particles. This is the idea of
Parker particle production (see [23–25]).
In the following, we will study Parker production of regular matter (i.e. ψ)
particles in our nonsingular bouncing cosmology. Particle production is particularly
efficient for modes which have a wavelength larger than the Hubble radius. For
such modes, the canonical field variable is frozen out and squeezed. If the squeezed
59
mode function which was initially positive frequency is expanded at late times into
plane wave modes, then it obtains a large effective negative frequency component
which corresponds to the production of potential χ particles (the use of the word
“potential” indicates that we can only speak of true particles once the modes have
reentered the Hubble radius and started to oscillate again). The mixing between
the initial positive frequency modes and the late time modes is given by Bogoliubov
coefficients, and their values give us the number of produced particles (see [55] for a
textbook treatment of quantum field theory in curved space-time).
We will focus on modes which have wavelengths which already at early times
(the time when we want to compute the density of regular matter particles) allow an
interpretation of the field excitations as particles (i.e. short wavelengths). We set up
these modes in their initial vacuum state and compute the Bogoliubov coefficients
at later times in order to determine the number of produced particles. Integrating
over Fourier modes then lets us calculate the energy density in the produced matter
particles.
To begin, let us recall the basic equations we will use. Consider a homogeneous,
isotropic and spatially flat FLRW universe. The corresponding background dynam-
ics is governed by the energy density and the pressure of the different scalar fields
(with energy densities are pressures denoted by ρ and P , respectively) through the
Friedmann equations, which are given by
1
H2 = (ρφ + ρψ ) ,
3Mp2
1
Ḣ = − (ρφ + Pφ + ρψ + Pψ ) , (4.10)
2Mp2
where H ≡ ȧ/a is the Hubble rate of the Universe and the dot denotes the derivative
with respect to the cosmic time. In addition, the scalar fields satisfy the continu-
ity equation so that covariant energy conservation is guaranteed. The continuity
equations can be derived by varying the Lagrangian with respect to the scalar fields.
Namely, if we focus on the matter field ψ with mass m, the resulting equation of
60
motion is
( + m2 )ψ = 0 , (4.11)
which is the Klein-Gordon equation for a massive scalar field minimally coupled to
gravity. We note that in the context of inflation, different couplings to gravity have
been explored, and the creation of massless scalar fields nearly conformally coupled
to gravity or massive scalar fields conformally coupled to gravity are more often
considered [31, 33, 36, 38].
At the background level, the matter field ψ is only a function of cosmic time.
However, in order to study the quantum dynamics of its particle generation, we
include the gradient term and hence write down the equation of motion of the field
as follows,
∇2
ψ̈ + 3H(t)ψ̇ − ψ + m2 ψ = 0 , (4.12)
a2
with ∇2 ≡ g ij ∇i ∇j . Note that, in the current model we have ignored possible
interactions between ψ and other fields. For now, we focus on particle production of
the field ψ only under gravitational effects. This assumption is in agreement with
our starting point that all fields are weakly coupled.
For convenience, let us work in conformal time, which is related to the cosmic
time via dη = a−1 (t)dt. As is well known, the matter equation simplifies if written
in terms of the rescaled field χ defined via
χ(η, x) = a(η)ψ(η, x) . (4.13)
The equation of motion for ψ is then easy to translate into an equation of motion for
χ. Since the equation of motion is linear, each Fourier mode evolves independently.
It is thus convenient to consider the Fourier transformation on the field χ,
d3 k
Z
χ(η, x) = χk (η)eik·x , (4.14)
(2π)3/2
61
and then track the dynamics of each Fourier mode along with the evolution of the
background universe. The equation of motion for a Fourier mode of χ becomes
χ00k + ωk2 (η)χk = 0 , (4.15)
with
a00
ωk2 ≡ k 2 + a2 m2 − , (4.16)
a
where primes denote derivatives with respect to conformal time. It is the final term
in (4.16) which is responsible for the squeezing of the modes.
4.3.1 Dynamics of the variable χk

We will solve the equation of motion (4.15) phase by phase throughout the cos-
mological evolution. Specifically, we divide the time line into four separate phases:
the initial matter-dominated phase of contraction, the Ekpyrotic phase of contrac-
tion, the nonsingular bouncing phase, and finally the fast roll phase of expansion. In
each of the phases, different approximations are applicable.
Matter contraction
We first study the dynamics of χk in the matter-dominated phase of contraction.
In this stage, the background scale factor evolves as
aE
a(η) ' cE (η − η̃E )2 , with cE ≡ , (4.17)
(ηE − η̃E )2
where we introduce aE ≡ a(ηE ) as the value of the scale factor at the end of this
phase (and correspondingly, at the beginning of the Ekpyrotic phase). The coeffi-
cient η̃E ≡ ηE − 2/HE would correspond to the conformal time at the Big Crunch
singularity if the matter phase of contraction lasted forever. Its inclusion is very
useful in regularizing the detailed computation. Additionally, we have introduced
the conformal Hubble rate H ≡ a0 /a.
62
Applying the relation (4.17) to the definition of the effective frequency square
(4.16), one can derive
2
ωkm (η)2 = k 2 + c2E m2 (η − η̃E )4 − , (4.18)
(η − η̃E )2
where the superscript “m” denotes the phase of matter contraction. Then, by solving
the key equation of motion (4.15), one can determine the solution of χk . At early
times, i.e. for η ηE , it is obvious that the first two terms in Eq. (4.18) for the
effective frequency are dominant. Accordingly, the solution for the variable χk is
approximately the plane wave solution. Requiring vacuum initial conditions, we find
the solution to be given by
m
e−iωk (η−η̃E )
χm
k (η) ' p m . (4.19)
2ωk
Ekpyrotic contraction
In the two-field matter bounce model, the potential of the scalar field φ is neg-
ative and of exponential form. This kind of potential can drive φ into the attractor
trajectory for which the effective equation of state is larger than unity [47]. As the
Universe contracts, the energy density of the scalar field φ catches up with that of
the matter field ψ, and subsequently, a period of Ekpyrotic contraction begins. Dur-
ing this period, the background scale factor can be expressed as a function of the
conformal time in the following approximate form
q aB−
a(η) ' cB− (η − η̃B− ) 1−q , with cB− ≡ q , (4.20)
(η − η̃B− ) 1−q
and where aB− ≡ a(ηB− ) is the scale factor at the end of Ekpyrotic phase of contrac-
tion (and equivalently, at the beginning of the bounce phase). Similar to the treat-
q
ment of the phase of matter contraction, we have introduced η̃B− ≡ ηB− − (1−q)H B−
63
in the above expression. Additionally, the parameter q is associated with the back-
ground dynamics and is required to be much less than unity so that the Ekpyrotic
phase is achieved1 .
Making use of (4.20), one can easily obtain the squeezing term in the equation
of motion for χk :
a00 q(1 − 2q) 1

=− . (4.21)
a (1 − q) (η − η̃B− )2
2
The mass term a2 m2 is subdominant since the scale factor has decreased to a small
value and the energy scale of the Universe has become higher than the mass scale m
after matter contraction. Therefore, the dominant part of Eq. (4.16) is given by
q(1 − 2q) 1
ωkc (η)2 = k 2 + , (4.22)
(1 − q) (η − η̃B− )2
2
where the superscript “c” denotes the phase of Ekpyrotic contraction.

Substituting the expression (4.22) into the equation of motion (4.15), one can
solve the differential equation and find the explicit general solution to be
p
χck (η) = η − η̃B− {C1 (k)Jνc [k (η − η̃B− )]
+C2 (k)Yνc [k (η − η̃B− )]} , (4.23)
where Jνc and Yνc are the two linearly independent Bessel functions of the first and
second kind, respectively. They have the same index
1 − 3q 1
νc = ∼ . (4.24)
2(1 − q) 2
1
One needs at least q < 1/3 in order to wash out anisotropies (i.e. w > 1), but
in general, the Ekpyrotic phase of contraction requires that q 1/3 so that the
equation of state parameter is w 1.
64
For small values of the argument, i.e. for |k(η− η̃B− )| < 1, the scalings of the solutions
are
χk ∼ (η − η̃B− )q and χk ∼ (η − η̃B− )1−q , (4.25)
which translates to
ψk ∼ const and ψk ∼ (η − η̃B− )1−2q . (4.26)
We see that there is one (almost) constant mode, while the second mode decays.
Modes in the large k limit with |k(η − η̃B− )| 1 are still inside the Hubble
radius and keep the oscillating behavior. In this case, the solution can be simplified
into the following asymptotic form
1
χck (η) ' √ C̄1 (k)eik(η−η̃B− ) + C̄2 (k)e−ik(η−η̃B− ) ,

(4.27)
2k
with
2i + qπ −2i + qπ
C̄1 = √ (C1 + iC2 ) , C̄2 = √ (C1 − iC2 ) , (4.28)
2 π 2 π
up to leading order. One can see that the C̄1 and C̄2 terms represent the negative
and positive frequency modes, respectively. If C̄1 is vanishing, the solution to χk is
identified as the vacuum fluctuation. However, if C̄1 is not zero, there will be particle
production from the vacuum state.
Bouncing phase
The phase of Ekpyrotic contraction stops when the scalar field φ evolves into the
ghost condensate state. Afterwards, the Universe experiences a nonsingular bounc-
ing phase. To quantitatively characterize this phase, it is useful to approximately
parametrize the Hubble parameter as a linear function of the cosmic time
H(t) = Υt , (4.29)
where Υ is a positive coefficient that determines how fast the bounce takes place. We
denote the beginning and the end of the bounce phase by tB− and tB+ , respectively.
65
Using conformal time, we can solve for the scale factor near the bounce,
1
a(η) = aB + a3B Υη 2 + O(Υ2 η 4 ) , (4.30)
2
under the fast bounce assumption (i.e. Υη 2 < 1 during the bouncing phase). We
note that it is convenient to set the bounce point at tB = ηB = 0. We can then write
down, up to leading order, the effective frequency square in the bounce phase,
ωkb (η)2 = k 2 + a2B (m2 − Υ) , (4.31)
where the superscript “b” denotes the bounce phase. In this case, the solution can
be simply expressed as follows,
1
χbk (η) ' p b D1 (k)eikη + D2 (k)e−ikη ,

(4.32)
2ωk
again for the large k modes.
Fast roll expansion

After the bounce, in the absence of backreaction of the fluctuations produced
before and during the bounce, the Universe would experience a period of fast roll
expansion with an effective equation of state w = 1. The scale factor then evolves as
1 aB+
a(t) = cB+ (η − η̃B+ ) 2 , with cB+ ≡ 1 . (4.33)
(ηB+ − η̃B+ ) 2
Analogously as before, aB+ is the value of the scale factor after the bounce phase
(and, correspondingly, at the beginning of this phase of fast roll expansion). Also,
we define η̃B+ ≡ ηB+ − 2H1B+ , which would correspond to the conformal time at the
big bang if there were no bounce. The above equation yields
a00 1
=− , (4.34)
a 4(η − η̃B+ )2
in the phase of fast roll expansion.
66
One can substitute the scale factor (4.33) and the derived relation (4.34) into
the expression (4.16) to determine the effective frequency square in the stage of fast
roll expansion. We find that the mass term is negligible when the Hubble parameter
of the Universe H is larger than the mass of the matter field m. As a result, we
obtain the dominant part of the effective frequency squared:
1
ωke (η)2 = k 2 + . (4.35)
4(η − η̃B+ )2
As usual, the superscript “e” denotes the phase of fast roll expansion.
Making use of (4.35), the equation of motion (4.15) yields the following solution,
p
χek (η) = η − η̃B+ {E1 (k)J0 [k(η − η̃B− )] + E2 (k)Y0 [k(η − η̃B− )]} , (4.36)
where E1 and E2 are two coefficients to be determined later. Again, we focus on the
regime of large k modes (|k(η − η̃B+ )| 1) and can get the simplified form,
1
χek (η) ' √ Ē1 (k)eik(η−η̃B+ ) + Ē2 (k)e−ik(η−η̃B+ ) ,

(4.37)
2k
with
2i + π/3
Ē1 = √ (E1 + iE2 ) ,
2 π
−2i + π/3
Ē2 = √ (E1 − iE2 ) , (4.38)
2 π
up to leading order.
4.3.2 Particle production

We are interested in whether particles of the matter field have been generated
in the process of the evolution during the above different phases. As is convention-
ally done in studying quantum field theory in curved space-time, we work in the
Heisenberg representation in which the operators carry the time dependence.
To study particle production, it is important to express the solution of the field
variable χk at a given time η as a sum of local positive and negative frequency modes
67
as follows,
1 Rη Rη
χk = √ [αk e−i ωk dη̃ + βk ei ωk dη̃ ] , (4.39)
2ωk
where the coefficients αk and βk are the so-called Bogoliubov coefficients. If the mode
functions evolve adiabatically, then the Bogoliubov coefficients are time independent
and no particles are produced.
Following the standard approach of canonical quantization, one can define the
conjugate momentum Πk for the field variable through
δS
Πk ≡ = χ0k . (4.40)
δχ0k
Note that the general solutions to the field variable in different phases have
been obtained in the expressions (4.19), (4.27), (4.32), and (4.37), respectively. In
each phase, there are two modes. Given initial conditions at the beginning of the
dynamics, the actual solution in each phase is the combination of the two fundamental
solutions in that phase which is obtained by demanding the solutions to the variable χ
and its conjugate momentum (and hence its time derivative) to be continuous across
the transition between the previous phase and the one under consideration. Given
the solution in any given phase, we can then determine the Bogoliubov coefficients.
In doing this, we interpret the “amount” of negative frequency modes that are created
during the entire evolution as the particle production.
We wish to consider the evolution of the vacuum state of χvac k (η) defined at some
initial time ηi . First of all, we expand the field into combinations χvac k and χvac∗
k of
the two fundamental solutions of the classical mode equation which are positive and
negative frequency at ηi , respectively. After quantization, the coefficients ak and
a†k become operators obeying the canonical commutation relations, which in turn
implies that the χ field and its conjugate momentum Π are canonically conjugate
operators. It is then possible to write the mode expansion of the χ field as
d3 k
Z
vac ik·x † vac∗ −ik·x
χ(x) = a k χ k e + a χ
k k e , (4.41)
(2π)3/2
68
where a†k and ak represent the creation and annihilation operators, respectively. The
time dependence of the field operator is manifested by the time dependence of the
classical mode functions.
At a later time ηnew we must expand the field operator χ in a new basis of mode
functions which are locally positive and negative frequency at the new time. Let us
denote the positive frequency mode at time ηnew by χnew k (η). The field χ can also be
expanded in terms of the new basis modes
d3 k
Z
new ik·x † new∗ −ik·x
χ(x) = b χ
k k e + b χ
k k e , (4.42)
(2π)3/2
where b†k and bk are new creation and annihilation operators. Since χvac k and χnew
k
form complete sets of mode solutions at early and late times, it is possible to relate
one set of modes to the other via
χnew
k = αk χvac vac∗
k + βk χk , (4.43)
where the coefficients are the famous Bogoliubov coefficients. From this, it is straight-
forward to derive that the expectation value of the vacuum number operator evalu-
ated after the phase transition is equal to
nk = |βk |2 . (4.44)
This is the result that we will exploit throughout this work and interpret as the
particle number in mode k which has been produced between the initial time and
the time ηnew .
Particle number at the beginning of the Ekpyrotic phase

We start our modes in their vacuum state during the matter phase of contraction.
At the time tE , the equation of state transits into an Ekpyrotic one. As we found in
the previous subsections, the fluctuating mode solutions χk in the matter contracting
phase and Ekpyrotic contraction period are given by (4.19) and (4.27), respectively.
To relate the two in the limit of small q, we can apply the matching condition
69
by requiring the solutions and their conjugate momenta to be continuous at the
transition point. This then yields up to leading order the following relation:
c m
e−iωk (ηE )ηE −iωk (ηE )(ηE −η̃E ) 2q(η̃B− + ηE )

2 2 4
βk ' √ − 10m cE (ηE − η̃E ) , (4.45)
8 3k 2 (η̃B− − ηE )3
for fluctuation modes with comoving wave number k larger than the comoving Hubble
parameter HE . The number of particles that is then produced is
40qm2 c2E (ηE − η̃E )4 (η̃B− + ηE )

2 1 4 4 8
nk = |βk | ' 100m cE (ηE − η̃E ) − .
192k 4 (η̃B− − ηE )3
(4.46)
To evaluate the contribution of these new particles to the background evolution,

one should compute the energy density of these particles and compare it to the
background energy density. Let us take for granted the fact that the energy density
of the particles that have been produced is given by
Z Z
1 1
ρχ = d k nk ωk ' 2 4 dk k 3 nk ,
3
(4.47)
(2π)3 a4 2π a
in the limit where ωk ' k. We will justify that this is the form that the energy
density of particle creation should take in Sec. 4.4. However, similar to the vacuum
energy density, it is easy to realize that this integral diverges, both in the ultraviolet
(UV) and in the infrared (IR). For this reason, it is necessary to introduce some
cutoffs. First, all of the above analysis has been done for modes on sub-Hubble
scales at the initial time. Thus, the comoving scale corresponding to the initial
Hubble radius yields the natural IR cutoff. Second, we impose an UV cutoff that
corresponds to the Hubble scale at the transition between matter contraction and
Ekpyrotic contraction. In other words, we integrate modes with value of k smaller
than HE . When considering particle production at the beginning of the bounce
phase, we need to consider modes with k between HE and HB− . During the bounce
phase, we take the Planck scale to be the UV cutoff, so we integrate modes with
scale between HB− and aB Mp (see Fig. 4–1 for a pictorial representation).
70
Using Eq. (4.46) for the number of particles created at tE , performing the integral
(4.47), and dividing the resulting energy density by the background energy density
ρback ' 3Mp2 HE2 , we find the following result for the energy density of the particles
produced during the matter phase of contraction (as a fraction of the background
energy density):
m2

ρχ 2 (η̃B− + ηE ) aB− HB−
' 30m − 7q ln , (4.48)
ρback 336π 2 Mp2 HE2 (η̃B− − ηE )3 kmin
where kmin represents the comoving wave number of the fluctuating mode with the
largest wavelength.
Let us explore the consequences of Eq. (4.48). Overall, the energy density of
particle production after the matter phase of contraction is very small compared
to the background energy density. The prefactor of Eq. (4.48) can be of order 1,
but the square bracket is suppressed due to the mass squared term and the second
term is suppressed due to the large duration of the Ekpyrotic phase. The latter is
necessary to dilute unwanted anisotropies (see [20]). Finally, the log term could be
arbitrarily large for arbitrarily small values of kmin , but recalling that the fluctuations
must reenter the Hubble radius after the bounce, this term cannot contribute to the
energy density significantly.
Particle number at the beginning of the bounce phase

Using the same technique as above, we require the mode solutions χk given by
Eqs. (4.27) and (4.32) to be continuous at the transition time tB− . The matching
conditions then yield the following Bogoliubov coefficient:
b ωkb − k
βk ' e−ik(ηB− −η̃E )−iωk ηB− p , (4.49)
2 3kωkb
where we recall that the effective frequency in the bounce phase is given by Eq.
(4.31). Up to leading order in k, we find the resulting number of particles to be
given by
a4 (m2 − Υ)2
nk = |βk |2 ' B . (4.50)
48k 4
71
Using the fact that m2 Υ and, since the bounce phase is short, assuming that
a(ηB− ) ' aB , the energy density of particle production is found to be
Υ2

aB− HB−
ρχ ' ln , (4.51)
96π 2 aE HE
and in terms of a fraction of the background energy density, we find

ρχ 1 aB− HB−
' ln . (4.52)
ρback 288π 2 Mp2 t2B− aE HE
Note that we used the relation HB− = ΥtB− .

This last result shows that gravitational particle production starts to become
important in the Ekpyrotic phase of contraction. If the bounce phase is very fast (a
few Planck times for instance), then the energy density of particle production is of
the order of O(10−4 ) times the background energy density. However, recent studies
suggest that the bounce phase should last about 103 − 104 Planck times to obtain
the correct amplitude of cosmological perturbations and a sufficiently small tensor-
to-scalar ratio [56, 57], which reduces the fraction (4.52).
Particle number at the end of the bounce phase

For the transition in the equation of state occurring at tB+ , we require continuity
of the mode solutions given by Eqs. (4.32) and (4.37). Using the same techniques as
before, we find the Bogoliubov coefficient βk and the resulting number of particles.
The latter is found to be
a4B Υ2
nk ' (4Υ2 t4B+ − Υt2B+ − 1)2 , (4.53)
48k 4
where we used the facts that m2 Υ, a(ηB+ ) ' aB , and ηB+ ' tB+ /aB+ . The
resulting energy density in terms of the background energy density is found to be
(4Υ2 t4B+ − Υt2B+ − 1)2

ρχ aB Mp
' ln . (4.54)
ρback 288π 2 Mp2 t2B+ aB− HB−
72
Here again, we see that particles have been produced. In fact, we see that Eqs.
(4.52) and (4.54) share the same denominator if one assumes that the bounce phase
is symmetric (|tB− | = |tB+ |). However, the energy density after the bounce phase
has an extra factor which tends to 1 as the bounce duration tends to 0. So, in
the short bounce limit, the energy densities of particle production at the end of the
Ekpyrotic phase and at the end of the bounce phase are comparable, whereas for a
longer bounce, the contribution from the bounce phase dominates. At this point,
the energy density of particle production is at most O(10−3 ) times the background
energy density.
4.3.3 Reheating time and reheating temperature

We found in the previous section that particles are gravitationally produced
throughout the evolution of the Universe, but that the dominant contribution came
from the bounce phase, and that the energy density of these particles could reach
about 10−3 times the background energy density at the end of the bounce phase. On
one hand, this is not enough to disturb the background evolution, i.e., it does not
lead to large backreaction effects. On the other hand, this is enough to reheat the
Universe since after the bounce phase, particles that have been produced will redshift
less fast than the background field and will hence ultimately become dominant.
Specifically, we know that ρχ ∼ a−4 , whereas ρback ∼ a−6 . Defining the “reheat-
ing time” tR as the time when the total energy density of particle production equals
the background energy density,
ρχ,total (tR ) = ρback (tR ) , (4.55)
we find that reheating occurs at
Mp3 a6B+
tR = t̃B+ + , (4.56)
(3ρ∗ )3/2 (tB+ − t̃B+ )2
73
where
Υ2 a4B+

aB Mp
ρ∗ = (4Υ2 t4B+ − Υt2B+ − 1)2 ln (4.57)
96π 2 aB− HB−
is the energy density of particle production at the end of the bounce phase such that
ρχ,total (tB+ ) = ρ∗ a−4 (tB+ ) . (4.58)
Finally, using the relation

π2 4
T ,
ρχ,total (tR ) = (4.59)
15 R
we find that the reheating temperature is given by
1/4 3/4
15 ρ∗
TR = (4.60)
9 π Mp HB+ a3B+
1/2
3/4
101/4 Υ3/2

2 4 2 2 aB Mp
= (4Υ tB+ − ΥtB+ − 1) ln .
48π 2 Mp HB+ aB− HB−
For a short bounce, this temperature can be of the order of O(10−5 )Mp , about an
order of magnitude below the GUT scale. Using parameter values that better suit
the recent observations [56], we find TR ∼ O(10−7 )Mp .
Let us briefly compare reheating in inflationary cosmology and the bounce re-
heating mechanism by Parker particle production discussed here. Neither Parker
particle production nor inflationary preheating produce a state with a thermal dis-
tribution of particles at the time that matter starts to dominate the energy density.
A period of thermalization (e.g. via perturbative processes) is required both in infla-
tion and in our bouncing cosmology. In fact, the processes involved will be identical.
If we compare the energy density at which regular matter starts to dominate, then
the value TR ∼ O(10−7 )Mp in our scenario is comparable to the “temperature” after
preheating in intermediate energy inflation models. For a review of the challenges of
the actual thermalization process the reader is referred to [1].
74
4.4 Backreaction of Thermal Particles from Cosmic Fluctuations
In this section we take another view on the backreaction of the produced particles
on the background dynamics. Instead of focusing directly on the number of produced
particles and computing their associated energy density we will consider the energy-
momentum tensor of the scalar field and directly compute its backreaction.
We know that for a scalar field ψ with a canonical Lagrangian given by Eq.
(4.8), the energy density due to modes with k H is given by
1 1 1
ρbr ' h(δ ψ̇)2 i + 2 h(∇δψ)2 i + m2 h(δψ)2 i , (4.61)
2 2a 2
where δψ denotes the fluctuation of the scalar field. Note that in contrast to a
field φ responsible for Ekpyrotic contraction, ψ has no homogeneous nonvanishing
background value. In the above expectation values, only sub-Hubble Fourier modes
are considered. If we wanted to discuss the backreaction of super-Hubble modes, it
would be essential to include the metric fluctuations induced by the matter modes. In
the context of inflationary cosmology, the backreaction formalism for long wavelength
modes was developed in detail in [58].
Recalling that χ = aψ and using the mode expansion of the χ field, it is straight-
forward to express the above energy density as
Z
1 3 1

0 2 0∗ 0 ∗ 2 2 2 2
2

ρbr ' d k |χ k | − H (χ k kχ + χ χ
k k ) + H + k + m a |χ k | .
(2π)3 a4 2
(4.62)
Expanding the χk modes as sums of positive and negative frequency modes [see Eq.
(4.39)] and using the fact that ωk ' k in the large k limit, the backreaction energy
density becomes Z
1 1
d3 k ωk |αk |2 + |βk |2 .

ρbr ' 3 4
(4.63)
(2π) a 2
Noting that |αk |2 − |βk |2 = 1 and nk = |βk |2 , we can reexpress the last result as
Z Z
1 3 1 1
ρbr ' d k ωk + d3 k ω k nk . (4.64)
(2π)3 a4 2 (2π)3 a4
75
The first term of this equation is the (divergent) vacuum energy, which we ignore.
The second term is what we wished to show as being the energy density of particle
production and it is what we used throughout Sec. 4.3.2.
4.5 Particle Production Through Direct Interactions

In inflationary cosmology particle production through direct interactions be-
tween the matter field and the inflaton field is much more important than gravi-
tational particle production. The energy density due to gravitational particle pro-
duction is of the order H 4 , where H is the Hubble expansion rate during inflation,
whereas the energy density transferred to matter via preheating is proportional to
H 2 Mp2 which is parametrically larger by a factor of (Mp /H)2 . Thus, an obvious ques-
tion to ask is how large the contribution of direct interactions to matter production
is in our matter bounce scenario.
The efficiency of the parametric resonance instability [2] underlying preheating
can be traced [5] to the fact that the inflaton oscillates many times through ϕ = 0.
During each crossing, the adiabaticity condition for the matter modes is violated,
leading to many bursts of particle production. In contrast, in our model φ crosses
zero only once, and so less particle production through direct interactions is expected.
On the other hand, in the emergent Galileon cosmology of [8], particle produc-
tion via direct interactions turns out to be very efficient in spite of the fact that the
dynamical background field crosses φ = 0 only once [9]. However, as explained in [9],
this is unexpected and due to the particular couplings of the model. Hence, taking
the lessons of [9] into account we expect a small amount of particle production via
interactions. In the following, we will verify this expectation.
To obtain direct particle production we need to add an interaction Lagrangian
to the problem (as one needs to in order to obtain inflationary reheating). We take
this coupling to be given by the last term in the following Lagrangian
1
L = K(φ, X) + G(X)φ + P (ψ, Y ) − λ2 φ2 ψ 2 Mp2 , (4.65)
2
76
for some real coupling constant λ. We assume that the coupling only takes place
during the bounce phase where we expect particle production to be most significant.
Outside the bounce phase, we assume that the coupling constant λ is very small 2 .
Using the above Lagrangian with the interaction term, the equation of motion
(EoM) for ψ in Fourier space becomes
2
k 2 2 2 2
ψ̈k + 3H ψ̇k + + m + λ φ Mp ψk = 0 . (4.66)
a2
Defining a new auxiliary field as Xk (t) ≡ a3/2 (t)ψk (t), the equation of motion sim-
plifies and becomes
Ẍk + ωk2 Xk = 0 , (4.67)
where the time-dependent effective frequency is given by
k2 9 2 3
ωk2 = + m 2
+ λ2 2
φ Mp
2
− H − Ḣ . (4.68)
a2 4 2
Since the background evolution in the bounce phase is well known (see [20]), we
could write down the form of Eq. (4.68) explicitly, but we will be mainly interested
in the effective frequency close to the bounce point where φ crosses 0. We note that
√
π t
φ(t) ' φ̇B T erf , (4.69)
2 T
where T is given by s
HB+ 2
T ' . (4.70)
Υ ln(φ̇2B /6HB+
2
)
2
In other words, we take λ to be a time-dependent coupling.
77
Thus, as long as the ratio t/T remains small, it is sufficient to expend the effective
frequency squared to second order in time as follows,
2
9Υ2 k 2 Υ 2

2 k 2 3Υ 2 2 2
ωk (t) ' +m − + λ φ̇B Mp − − 2 t (4.71)
a2B 2 4 aB
≡ b + ct2 ,
where we have denoted the zeroth-order term as b and the coefficient of the quadratic
term as c.
To determine when particle production occurs, we need to find when the adia-
baticity condition is violated. This is the case when
|ω̇k | & ωk2 . (4.72)
Using Eq. (4.71) for the effective frequency, the inequality (4.72) is equivalent to
c3 t6 + 3bc2 t4 + c(3b2 − c)t2 + b3 . 0 . (4.73)
Since the constant term in the above equation is negligible when λ is large, we can
solve for t and thus we find that particle production occurs without interruption for
ti . t . tf with
√
s
4c − 3b2 − 3b
tf = (4.74)
2c
and ti = −tf . We note that this is only valid if c > 3b2 , which translates into an
upper bound3 , v
u q
u 4Υ + 3λ2 φ̇2 M 2 − 11Υ2
t B p
kmax = aB , (4.75)
3
as long as λ > 3Υ/φ̇B Mp .
3
As in Sec. 4.3.2, we assume m2 Υ.
78
In order to solve the EoM, let us write down the solution in the form of 4
1 h
−i t dt̃ ω (t̃)
R
i t dt̃ ω (t̃)
R i
Xk (t) = p αk (t)e ti k + βk (t)e ti k . (4.76)
2ωk (t)
Its conjugate momentum is then given by

r
ωk (t) h −i tt dt̃ ωk (t̃)
R
i tt dt̃ ωk (t̃)
R i
Πk (t) = i −αk (t)e i + βk (t)e i , (4.77)
2
and the Bogoliubov coefficients satisfy the following equations:
ω̇k 2i Rtt dt̃ ωk (t̃)
α̇k = e i βk , (4.78)
2ωk
ω̇k −2i Rtt dt̃ ωk (t̃)
β̇k = e i αk . (4.79)
2ωk
To set the initial conditions, we require that there are no particles at ti , which implies
that αk (ti ) = 1 and βk (ti ) = 0.
Once we find the solution Xk (t), we know that the total number of particles
that will have been produced is given by
!
ωk (tf ) |Ẋk (tf )|2 2 1
nk (tf ) = 2
+ |Xk (tf )| − , (4.80)
2 ωk (tf ) 2
or equivalently by
nk (tf ) = |βk (tf )|2 . (4.81)
An approximate solution is given by (see the Appendix for details)

2√
tf c ct2f

2 b
nk (tf ) ' sinh − √ ln 1 + . (4.82)
2 2 c b
4
We use the WKB approximation.
79
The energy density of particle production at the end of the bounce phase is then
given by Z kmax
1
ρX (tB+ ) = 2 dk k 2 ωk (tB+ )nk (tf ) , (4.83)
2π a(tB+ )4 0
where we used the fact that nk (tB+ ) = nk (tf ) since particle production stops before
the end of the bounce phase. As explained in the Appendix, the result for nk (tf )
p
is only valid for b > 0, which imposes a lower bound kmin = aB 3Υ/2. Since we
are interested in the limit where λ Υ, we notice that kmax kmin meaning that
the integral is dominated by its UV cutoff. Thus, we can effectively take the lower
integration bound to be 0.
The integrand of Eq. (4.83) is a complicated function of k, but it is straight-
forward to find an upper bound to the overall contribution. For large values of the
coupling constant5 , we see from Eq. (4.68) that the effective frequency squared at
the end of the bounce phase is dominated by the term λ2 φ(tB+ )2 Mp2 . We also notice
that nk (tf ) . sinh2 (1/2) for all allowed values of k, so Eq. (4.83) simplifies to become
the inequality
a3B (λφ̇B Mp )5/2 HB+ erf(tB+ /T ) sinh2 (1/2)

ρX (tB+ ) . q . (4.84)
7/4 3/2 4 2 2
3 (2π) aB+ Υ ln(φ̇B /6HB+ )
5
One may worry that a large coupling constant leads to a Landau pole by renor-
malization group running to the UV. At this point, we only want to see the limitations
of particle production via this coupling, hence we try to find an upper bound to the
overall contribution and we consider a large coupling constant as a limit. However,
we do not really expect this coupling constant to be large in nature, and so, a Landau
pole should be safely avoided.
80
This can be compared to the energy density of Parker particle production (recall Eq.
(4.54)),
√
ρX (tB+ ) 8 2πa3B (λφ̇B Mp )5/2 tB+ erf(tB+ /T ) sinh2 (1/2)
ρχ (tB+ )
. r 2 . (4.85)
a M φ̇
33/4 a4B+ Υ2 (4Υ2 t4B+ − ΥtB+ − 1)2 ln aB−B HB−p
ln 6HB2
B+
From this result, we see that for a fixed energy scale HB+ and a fixed coupling con-
stant λ, the ratio of the energy densities tends to zero as Υ → 0 or as Υ → ∞. This
means that in the limit where the bounce is either infinitely long or infinitely short,
the energy density from Parker particle production will always dominate the energy
density from interaction. However, for typical bounce durations, the ratio of the
energy densities depends on the size of the coupling constant. An example is shown
in Fig. 4–2 where we see that particle production from interaction can contribute
significantly to Parker particle production, but only if the coupling between the two
fields is quite large. In general, for smaller values of the coupling constant, the en-
ergy density from Parker particle production dominates over the energy density from
interaction.
4.6 Conclusions
We have considered Parker particle production in the two-field matter bounce
model of [19], a prototypical example of a nonsingular matter bounce which can
generate a scale-invariant spectrum of cosmological perturbations and which is stable
against anisotropies near the bounce point. The stability to anisotropies is obtained
via an Ekpyrotic phase of contraction. This phase of Ekpyrosis, on the other hand,
also washes out any matter and radiation which might have existed early in the
contracting phase. Hence, to make the model viable, a reheating mechanism is
required.
We have shown that in our background, gravitational Parker particle produc-
tion is sufficiently effective to reheat the Universe to high temperatures. For minimal
coupling between the scalar field φ yielding the Ekpyrotic contraction and the nonsin-
gular bounce and regular matter the effect of particle production through interactions
81
−2 Numerical Computation
Analytic Upper Bound
Parker Particle Production
−4
log10[ρ/ρback] −6
−8
−10
−12
10−4 10−3 10−2 10−1
λ
Figure 4–2: Plot of the ratio of the energy density of particle production to the
background energy density as a function of the coupling constant. The slope of
the Hubble parameter and the physical duration of the bounce are taken to be
Υ = 5.5 × 10−7 Mp2 and tB+ = 2.1 × 103 Mp−1 , respectively. The blue curve shows
the analytic upper bound found in Eq. (4.84) and the black curve was obtained by
solving the EoM and the integral for the energy density numerically. The dashed red
line is the contribution from Parker particle production that we found earlier [see
Eq. (4.54)].
is generally small, but could contribute to Parker particle production if the coupling
were large. We thus see that we do not need to introduce extra ingredients into our
matter bounce model in order to obtain a hot postbounce universe dominated by
regular matter and radiation. This contrast with the case of inflationary cosmology,
where the direct coupling of the inflaton field with matter generically produces an
energy density in the reheat fields which is of the order of the inflationary energy
density, whereas the contribution of Parker particle production is of the order ρ ∼ H 4
and which is hence suppressed compared to the density of particles produced by pre-
heating by a factor of order (H/Mp )2 . The reason for this difference is related to the
fact that in the case of inflation the inflaton field loses most of its energy density
82
during the reheating process, whereas in our nonsingular bounce the field φ retains
most of its energy.
Finally, we note that the methods developed here are applicable to the study
of Parker particle production in other nonsingular bouncing models, e.g. in the new
Ekpyrotic universe [59]. Based on the arguments of the previous paragraph we con-
jecture that also in these other models, Parker particle production will be sufficient
to reheat the Universe, and that Parker particle production will not be suppressed
relative to particle production via direct couplings. To support this conjecture, we
point out that Parker particle production has been shown to be efficient for certain
models of the matter bounce in loop quantum cosmology [60].
83
ACKNOWLEDGEMENTS
We thank Jaume Haro for useful discussions. JQ acknowledges the Natural

Sciences and Engineering Research Council (NSERC) of Canada for financial support
under a CGS M scholarship. The research of RB and YFC is supported by an NSERC
Discovery grant and by funds from the Canada Research Chair program.
84
Solving for the Bogoliubov coefficients in the bounce phase
We want to solve the set of coupled first-order ordinary differential equations
ω̇k 2i Rtt dt̃ ωk (t̃)

α̇k = e i βk , (4.86)
2ωk
ω̇k −2i Rtt dt̃ ωk (t̃)
β̇k = e i αk . (4.87)
2ωk
To simplify the treatment, let us define
ω̇k ±2i Rtt dt̃ ωk (t̃)

gk± (t) ≡ e i , (4.88)
2ωk
together with the matrix

 
0 gk+ (t)
Ak (t) ≡  (4.89)
 

gk− (t) 0
and the vector  

 αk (t) 
yk (t) ≡   . (4.90)
βk (t)
Then, the system to solve simply reads
ẏk (t) = Ak (t)yk (t) , (4.91)
and the initial conditions αk (ti ) = 1 and βk (ti ) = 0 can be written as

 
 1 
yk (ti ) =   ≡ y0 . (4.92)
0
85
We make use of the Magnus expansion [61], which tells that the solution to the above
initial value problem is given by
yk (t) = eΩk,1 (t)+Ωk,2 (t)+... y0 , (4.93)
where
Z t
Ωk,1 (t) = dt1 Ak (t1 ) , (4.94)
ti
1 t
Z Z t1
Ωk,2 (t) = dt1 dt2 [Ak (t1 ), Ak (t2 )] , (4.95)
2 ti ti
and so on. Here, we consider the first-order approximation only. Defining

Z t
Ik± (t) ≡ dt1 gk± (t1 ) , (4.96)
ti
the solution becomes

 
0 Ik+ (t)
yk (t) ' exp   y0 , (4.97)
 
Ik− (t) 0
and therefore the Bogoliubov coefficients are given by

q
αk (t) ' cosh Ik+ (t)Ik− (t) , (4.98)
s
Ik− (t)
q
βk (t) ' sinh Ik+ (t)Ik− (t) . (4.99)
Ik+ (t)
In the present case for ωk (t) ' b + ct2 , the function gk± (t) defined in Eq. (4.88) can
be evaluated exactly. Then, under the large-λ assumption and as long as b > 0, the
integral defined in Eq. (4.96) evaluated at the end of the particle production phase
becomes
86
√
±itf b+ct2f
ct2f

± ie 2
Ik (tf ) ' ∓ √ ctf − b ln 1 + . (4.100)
2 c b
Therefore, the number of particles that has been produced is found to be
ct2f

2 2 1 2
nk (tf ) = |βk (tf )| ' sinh √ ctf − b ln 1 + . (4.101)
2 c b
87
Bibliography
[1] R. Allahverdi, R. Brandenberger, F. -Y. Cyr-Racine and A. Mazumdar, Reheat-
ing in Inflationary Cosmology: Theory and Applications, Ann. Rev. Nucl. Part.
Sci. 60, 27 (2010) [arXiv:1001.2600 [hep-th]].
[2] J. H. Traschen and R. H. Brandenberger, Particle Production During Out-of-
equilibrium Phase Transitions, Phys. Rev. D 42, 2491 (1990).
[3] L. Kofman, A. D. Linde and A. A. Starobinsky, Reheating after inflation, Phys.
Rev. Lett. 73, 3195 (1994) [hep-th/9405187].
[4] Y. Shtanov, J. H. Traschen and R. H. Brandenberger, Universe reheating after
inflation, Phys. Rev. D 51, 5438 (1995) [hep-ph/9407247].
[5] L. Kofman, A. D. Linde and A. A. Starobinsky, Towards the theory of reheating
after inflation, Phys. Rev. D 56, 3258 (1997) [hep-ph/9704452].
[6] R. H. Brandenberger and C. Vafa, Superstrings in the Early Universe, Nucl.
Phys. B 316, 391 (1989);
A. Nayeri, R. H. Brandenberger and C. Vafa, Producing a scale-invariant spec-
trum of perturbations in a Hagedorn phase of string cosmology, Phys. Rev. Lett.
97, 021302 (2006) [hep-th/0511140].
[7] R. H. Brandenberger, String Gas Cosmology, String Cosmology, J.Erdmenger
(Editor). Wiley, 2009. p.193-230 [arXiv:0808.0746 [hep-th]];
T. Battefeld and S. Watson, String gas cosmology, Rev. Mod. Phys. 78, 435
(2006) [hep-th/0510022].
[8] P. Creminelli, A. Nicolis and E. Trincherini, Galilean Genesis: An Alternative
to inflation, JCAP 1011, 021 (2010) [arXiv:1007.0027 [hep-th]].
[9] L. Perreault Levasseur, R. Brandenberger and A. -C. Davis, Defrosting in an
Emergent Galileon Cosmology, Phys. Rev. D 84, 103512 (2011) [arXiv:1105.5649
[astro-ph.CO]].
88
89

D 60, 023507 (1999) [arXiv:gr-qc/9809062].
[11] F. Finelli and R. Brandenberger, On the generation of a scale-invariant spectrum
of adiabatic fluctuations in cosmological models with a contracting phase, Phys.
Rev. D 65, 103522 (2002) [arXiv:hep-th/0112249].
[12] Y. F. Cai, T. Qiu, Y. S. Piao, M. Li and X. Zhang, Bouncing Universe with
Quintom Matter, JHEP 0710, 071 (2007) [arXiv:0704.1090 [gr-qc]].
[13] C. Lin, R. H. Brandenberger and L. Perreault Levasseur, A Matter Bounce By
Means of Ghost Condensation, JCAP 1104, 019 (2011) [arXiv:1007.2654 [hep-
th]].
[14] C. Kounnas, H. Partouche and N. Toumbas, Thermal duality and non-singular
cosmology in d-dimensional superstrings, Nucl. Phys. B 855, 280 (2012)
[arXiv:1106.0946 [hep-th]];
I. Florakis, C. Kounnas, H. Partouche and N. Toumbas, Non-singular string
cosmology in a 2d Hybrid model, Nucl. Phys. B 844, 89 (2011) [arXiv:1008.5129
[hep-th]];
R. H. Brandenberger, C. Kounnas, H. Partouche, S. P. Patil and N. Toum-
bas, Cosmological Perturbations Across an S-brane, JCAP 1403, 015 (2014)
[arXiv:1312.2524 [hep-th]].
80, 043516 (2009) [arXiv:0904.2835 [hep-th]].
[16] T. Biswas, A. Mazumdar and W. Siegel, Bouncing universes in string-inspired
gravity, JCAP 0603, 009 (2006) [arXiv:hep-th/0508194].
JCAP 1303, 026 (2013) [arXiv:1211.6269 [gr-qc]];
Y. -F. Cai and E. Wilson-Ewing, Non-singular bounce scenarios in loop quan-
tum cosmology and the effective field description, JCAP 1403, 026 (2014)
[arXiv:1402.3009 [gr-qc]].
[18] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, Oscillatory approach to a
singular point in the relativistic cosmology, Adv. Phys. 19, 525 (1970).
[19] Y. -F. Cai, D. A. Easson and R. Brandenberger, Towards a Nonsingular Bounc-
ing Cosmology, JCAP 1208, 020 (2012) [arXiv:1206.2382 [hep-th]].
90
[20] Y. -F. Cai, E. McDonough, F. Duplessis and R. H. Brandenberger, Two Field

[21] Y. -F. Cai, R. Brandenberger and P. Peter, Anisotropy in a Nonsingular Bounce,
Class. Quant. Grav. 30, 075019 (2013) [arXiv:1301.4703 [gr-qc]].
[22] Y. -F. Cai, Exploring Bouncing Cosmologies with Cosmological Surveys, Sci.
China Phys. Mech. Astron. 57, 1414 (2014) [arXiv:1405.1369 [hep-th]].
[23] L. Parker, Particle creation in expanding universes, Phys. Rev. Lett. 21, 562
(1968).
[24] L. Parker, Quantized fields and particle creation in expanding universes. 1.,
Phys. Rev. 183, 1057 (1969).
[25] L. Parker, Quantized fields and particle creation in expanding universes. 2.,
Phys. Rev. D 3, 346 (1971) [Erratum-ibid. D 3, 2546 (1971)].
[26] T. Imamura, Quantized Meson Field in a Classical Gravitational Field, Phys.
Rev. 118, 1430 (1960).
[27] A. A. Grib and S. G. Mamaev, On field theory in the friedman space, Yad. Fiz.
10, 1276 (1969) [Sov. J. Nucl. Phys. 10, 722 (1970)].
[28] A. A. Grib, S. G. Mamaev and V. M. Mostepanenko, Particle Creation from
Vacuum in Homogeneous Isotropic Models of the Universe, Gen. Rel. Grav. 7,
535 (1976).
[29] Y. B. Zeldovich and A. A. Starobinsky, Particle production and vacuum polar-
ization in an anisotropic gravitational field, Sov. Phys. JETP 34, 1159 (1972)
[Zh. Eksp. Teor. Fiz. 61, 2161 (1971)].
[30] J. A. Frieman, Particle Creation in Inhomogeneous Space-times, Phys. Rev. D
39, 389 (1989).
[31] A. A. Starobinsky, A New Type of Isotropic Cosmological Models Without Sin-
gularity, Phys. Lett. B 91, 99 (1980).
[32] N. P. Myhrvold, Runaway Particle Production in De Sitter Space, Phys. Rev.
D 28, 2439 (1983).
[33] L. H. Ford, Gravitational Particle Creation and Inflation, Phys. Rev. D 35, 2955
(1987).
91
[34] C. Pathinayake and L. H. Ford, Particle Creation By A Selfcoupled Scalar Field,

Phys. Rev. D 35, 3709 (1987).
[35] T. Mishima and A. Nakayama, Particle Production in De Sitter Space-time,
Prog. Theor. Phys. 77, 218 (1987).
[36] B. Spokoiny, Deflationary universe scenario, Phys. Lett. B 315, 40 (1993) [gr-
qc/9306008].
[37] S. Biswas, J. Guha and N. G. Sarkar, Particle production in de Sitter space,
Class. Quant. Grav. 12, 1591 (1995).
[38] P. J. E. Peebles and A. Vilenkin, Quintessential inflation, Phys. Rev. D 59,
063505 (1999) [astro-ph/9810509].
[39] N. G. Sarkar and S. Biswas, Particle production in de Sitter space-time, Int. J.
Mod. Phys. A 15, 497 (2000).
[40] G. N. Felder, L. Kofman and A. D. Linde, Gravitational particle production and
the moduli problem, JHEP 0002, 027 (2000) [hep-ph/9909508].
[41] J. de Haro and E. Elizalde, Gravitational particle production in massive
chaotic inflation and the moduli problem, Phys. Rev. Lett. 108, 061303 (2012)
[arXiv:1201.1227 [gr-qc]].
[42] L. Parker, Particle creation and particle number in an expanding universe, J.
Phys. A 45, 374023 (2012) [arXiv:1205.5616 [astro-ph.CO]].
[43] J. Haro, Different interpretations of the particle production in quantum fields
theory, Int. J. Theor. Phys. 48, 825 (2009).
[44] J. Haro, Gravitational particle production: A mathematical treatment, J. Phys.
A 44, 205401 (2011).
[45] C. Armendariz-Picon, V. F. Mukhanov and P. J. Steinhardt, Essentials of k
essence, Phys. Rev. D 63, 103510 (2001) [astro-ph/0006373].
[46] G. W. Horndeski, Second-order scalar-tensor field equations in a four-
dimensional space, Int. J. Theor. Phys. 10, 363 (1974);
A. Nicolis, R. Rattazzi and E. Trincherini, The Galileon as a local modification
of gravity, Phys. Rev. D 79, 064036 (2009) [arXiv:0811.2197 [hep-th]].
92
[47] J. Khoury, B. A. Ovrut, P. J. Steinhardt and N. Turok, The ekpyrotic universe:

Colliding branes and the origin of the hot big bang, Phys. Rev. D 64, 123522
(2001) [arXiv:hep-th/0103239];
J. Khoury, B. A. Ovrut, N. Seiberg, P. J. Steinhardt and N. Turok, From big
crunch to big bang, Phys. Rev. D 65, 086007 (2002) [arXiv:hep-th/0108187].
[48] J. K. Erickson, D. H. Wesley, P. J. Steinhardt and N. Turok, Kasner and mix-
master behavior in universes with equation of state w ≥ 1, Phys. Rev. D 69,
063514 (2004) [hep-th/0312009].
[49] N. Arkani-Hamed, H. -C. Cheng, M. A. Luty and S. Mukohyama, Ghost conden-
sation and a consistent infrared modification of gravity, JHEP 0405, 074 (2004)
[hep-th/0312099].
[50] A. D. Dolgov and A. D. Linde, Baryon Asymmetry in Inflationary Universe,
Phys. Lett. B 116, 329 (1982).
[51] L. F. Abbott, E. Farhi and M. B. Wise, Particle Production in the New Infla-
tionary Cosmology, Phys. Lett. B 117, 29 (1982).
[52] A. D. Dolgov and D. P. Kirilova, On Particle Creation By A Time Dependent
Scalar Field, Sov. J. Nucl. Phys. 51, 172 (1990) [Yad. Fiz. 51, 273 (1990)].
[53] Y. -F. Cai, R. Brandenberger and X. Zhang, Preheating a bouncing universe,
Phys. Lett. B 703, 25 (2011) [arXiv:1105.4286 [hep-th]].
[54] P. Peter and N. Pinto-Neto, Primordial perturbations in a non singular bouncing
universe model, Phys. Rev. D 66, 063509 (2002) [hep-th/0203013];
P. Peter, N. Pinto-Neto and D. A. Gonzalez, Adiabatic and entropy perturbations
propagation in a bouncing universe, JCAP 0312, 003 (2003) [hep-th/0306005];
J. Martin and P. Peter, Parametric amplification of metric fluctuations through
a bouncing phase, Phys. Rev. D 68, 103517 (2003) [hep-th/0307077];
S. Alexander, T. Biswas and R. H. Brandenberger, On the Transfer of Adiabatic
Fluctuations through a Nonsingular Cosmological Bounce, arXiv:0707.4679 [hep-
th];
R. Brandenberger, H. Firouzjahi and O. Saremi, Cosmological Perturbations on
a Bouncing Brane, JCAP 0711, 028 (2007) [arXiv:0707.4181 [hep-th]];
Y. F. Cai, T. Qiu, R. Brandenberger, Y. S. Piao and X. Zhang, On Perturbations
of Quintom Bounce, JCAP 0803, 013 (2008) [arXiv:0711.2187 [hep-th]];
Y. -F. Cai, T. -t. Qiu, R. Brandenberger and X. -m. Zhang, A Nonsingular
93
Cosmology with a Scale-Invariant Spectrum of Cosmological Perturbations from

Lee-Wick Theory, Phys. Rev. D 80, 023511 (2009) [arXiv:0810.4677 [hep-th]];
X. Gao, Y. Wang, W. Xue and R. Brandenberger, Fluctuations in a Hořava-
Lifshitz Bouncing Cosmology, JCAP 1002, 020 (2010) [arXiv:0911.3196 [hep-
th]];
F. Finelli, P. Peter and N. Pinto-Neto, Spectra of primordial fluctuations in two-
perfect-fluid regular bounces, Phys. Rev. D 77, 103508 (2008) [arXiv:0709.3074
[gr-qc]].
[55] N. Birrell and P. Davies, Quantum Fields in Curved Space (Cambridge Univ.
Press, Cambridge, 1982).
[astro-ph.CO]].
[57] L. Battarra, M. Koehn, J. L. Lehners and B. A. Ovrut, Cosmological Perturba-
tions Through a Non-Singular Ghost-Condensate/Galileon Bounce, JCAP 1407,
007 (2014) [arXiv:1404.5067 [hep-th]].
[58] L. R. W. Abramo, R. H. Brandenberger and V. F. Mukhanov, The Energy -
momentum tensor for cosmological perturbations, Phys. Rev. D 56, 3248 (1997)
[gr-qc/9704037].
[59] A. Notari and A. Riotto, Isocurvature perturbations in the ekpyrotic universe,
Nucl. Phys. B 644, 371 (2002) [hep-th/0205019];
F. Finelli, Assisted contraction, Phys. Lett. B 545, 1 (2002) [hep-th/0206112];
E. I. Buchbinder, J. Khoury and B. A. Ovrut, New Ekpyrotic cosmology, Phys.
Rev. D 76, 123503 (2007) [hep-th/0702154];
P. Creminelli and L. Senatore, A Smooth bouncing cosmology with scale invari-
ant spectrum, JCAP 0711, 010 (2007) [hep-th/0702165];
J. -L. Lehners, P. McFadden, N. Turok and P. J. Steinhardt, Generating ekpy-
rotic curvature perturbations before the big bang, Phys. Rev. D 76, 103501 (2007)
[hep-th/0702153 [HEP-TH]].
[60] J. Haro and J. Amorós, Viability of the matter bounce scenario in F (T ) gravity
and Loop Quantum Cosmology for general potentials, JCAP 1412, no. 12, 031
(2014) [arXiv:1406.0369 [gr-qc]].
94
[61] W. Magnus, On the exponential solution of differential equations for a linear

operator, Commun. Pure Appl. Math. 7, 649 (1954).
Chapter 5
Conclusions and Outlook
In this thesis, we presented two recent papers on the topic of nonsingular bounc-
ing cosmology. We started by reviewing the motivation, the different constructions,
and the current status of nonsingular bouncing cosmology as a possible theory of the
very early universe.
The first paper presented in this thesis explored the possibility for different
models of bouncing cosmology to produce a tensor-to-scalar ratio r of the order of
0.2 as it was suggested by the BICEP2 experiment in early 2014. Although the
early results of the BICEP2 experiment have been refuted by the Planck-BICEP2
joint analysis, and although the currently accepted bound on the tensor-to-scalar
ratio has been lowered to r < 0.12, our conclusions remain the same. Bouncing
cosmologies naturally produce large amounts of primordial gravitational waves with
r ∼ O(1). This is generic to bouncing models which use matter violating the Null
Energy Condition and bouncing models which use alternatives to General Relativity
to avoid the singularity. In general, the observed bound on the tensor-to-scalar ratio
puts constraints on the parameter models of the different bouncing cosmologies, and
the curvaton mechanism appears to be the most natural way to suppress the tensor-
to-scalar ratio.
The second paper presented in this thesis explored different reheating mech-
anisms for the matter-Ekpyrotic model of nonsingular bouncing cosmology. This
article showed how gravitational particle production could occur thanks to the non-
trivial evolution of the quantum fields through the bounce and how this could lead
to a high enough reheating temperature. We also explored the case where there
could be a coupling between the background bounce field and the matter field. In
general, we found that a large enough coupling could contribute to reheating in this
model, but gravitational particle production remains the most important source of
95
particle creation. Furthermore, gravitational particle production is a more natural
mechanism to reheat the universe since it does not require any extra ingredient in
the model Lagrangian.
At the end of the review section in Chapter 2, we mentioned a few issues that
could arise in the context of nonsingular bouncing cosmology. In particular, there
was the problem of the evolution of curvature perturbations and the problem of
the generation of primordial non-gaussianities through a nonsingular bounce. Ad-
dressing these issues is necessary for bouncing cosmology to remain a competitive
alternative to the theory of inflationary cosmology. Perhaps even more problematic,
there are only few bouncing cosmologies which can generate a red-tilted power spec-
trum of curvature perturbations. Since the observational constraints on the tilt are
much more precise that the ones on r and non-gaussianities, bouncing models with a
scale-invariant or blue-tilted power spectrum of curvature perturbations can hardly
compete with inflation. An avenue that may deserve further attention in the future
is realizing the matter bounce with “known” fluids such as dark matter and dark
energy, and in this context, exploring the different avenues for quantizing the initial
fluctuations.
96

Topics in Bouncing Cosmology PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Topics in Bouncing Cosmology PDF

Uploaded by

Copyright:

Available Formats

Topics in Bouncing Cosmology

A thesis submitted to McGill University in partial fulfillment of the requirements of

I would like to express my deep and sincere gratitude to my supervisor Pro-

• J. Quintin, Y. F. Cai and R. H. Brandenberger, Matter creation in a nonsingu-

Papers not included in this Thesis

• Y. F. Cai, F. Chen, E. G. M. Ferreira and J. Quintin, A new model of axion

This thesis concerns two questions in bouncing cosmology, as addressed in the

Cette thèse concerne deux questions sur la théorie de l’univers rebondissant,

Dans la première partie de cette thèse, nous explorons la possibilité de produire

Dans la deuxième partie de cette thèse, nous explorons deux méchanismes de

2.1.1 The flatness problem

2.1.2 The horizon problem

2.1.3 The formation of structure problem

2.2 A short review of the theory of cosmological perturbations

Scalar metric fluctuations are generally represented by four degrees of freedom,

ds2 = a2 (η) [1 + 2Φ(η, x)] dη 2 − [1 − 2Φ(η, x)] dx2 ,

ds2 = a2 (η) dη 2 − [δij + hij (η, x)] dxi dxj

Equivalently for tensor modes, the Mukhanov variable is given by

where h is a polarization state of the tensor perturbation hij .

Equivalently, the equation of motion (EOM) for the curvature perturbation is

2.3 The generation of a scale-invariant power spectrum of curvature per-

Thus, we see that a scale-invariant (i.e. k-independent) power spectrum of curvature

L(φ, X) = (X − C 2 )2 − V (φ) , (2.35)

2.5 The matter-Ekpyrotic bounce

a(t) = (−t)p , (2.36)

where p ≡ 2/c2  1. This describes a contracting universe with equation of state

L = K(φ, X) + G(φ, X)φ , (2.39)

where we introduced X in Eq. (2.34) and where ≡ g µν ∇µ ∇ν is the d’Alembertian.

r < 0.12 (95% CL) . (2.47)

[26] J. Haro and J. Amorós, Matter Bounce Scenario in F(T) gravity,

[40] Y. F. Cai, T. Qiu, R. Brandenberger, Y. S. Piao and X. Zhang, On Perturbations

[54] V. A. Belinsky, I. M. Khalatnikov and E. M. Lifshitz, Oscillatory approach to a

[67] P. A. R. Ade et al. [BICEP2 Collaboration], Detection of B-Mode Polarization

3.2 Matter Bounce Cosmology

3.2.1 The matter bounce curvaton model

2.5 × 10−4 . m/Mp . 4.5 × 10−4 , (3.8)

3.2.2 New matter bounce cosmology

Fζ2 HE2 a2E 2 2

with γζ ' γψ and

Similarly to HE , we introduced aE , which is the value of the scale factor at the

Similarly to the previous subsection, we perform a numerical estimate to inves-

1.9 × 10−8 . HE /Mp . 1.9 × 10−6 , (3.15)

3.3 Implications for modified-gravity bouncing cosmology

3.3.1 Matter bounce in Hořava-Lifshitz gravity

3.3.2 The f (T ) matter bounce cosmology

3.3.3 Loop quantum cosmology

We are indebted to Robert Brandenberger and Jean-Luc Lehners for valuable

[10] R. H. Brandenberger, A. Nayeri, S. P. Patil and C. Vafa, Tensor Modes from a

[22] D. H. Lyth, BICEP2, the curvature perturbation and supersymmetry, JCAP

[34] M. W. Hossain, R. Myrzakulov, M. Sami and E. N. Saridakis, Evading Lyth

[46] Y. F. Cai, R. Brandenberger and P. Peter, Anisotropy in a Nonsingular

4.2 The Two-Field Matter Bounce Model

L = K(φ, X) + G(X)φ + P (ψ, Y ) , (4.1)

K(φ, X) = Mp2 [1 − g(φ)]X + βX 2 − V (φ) (4.4)

4.3 Gravitational particle production throughout the cosmic evolution

χ(η, x) = a(η)ψ(η, x) . (4.13)

χ00k + ωk2 (η)χk = 0 , (4.15)

4.3.1 Dynamics of the variable χk

a00 q(1 − 2q) 1

where the superscript “c” denotes the phase of Ekpyrotic contraction.

ψk ∼ const and ψk ∼ (η − η̃B− )1−2q . (4.26)

ωkb (η)2 = k 2 + a2B (m2 − Υ) , (4.31)

again for the large k modes.

where p ≡ 2/c2 1. This describes a contracting universe with equation of state