PhysRevD 101 045001

PHYSICAL REVIEW D 101, 045001 (2020)
Gravitational and chiral anomalies in the running vacuum universe

and matter-antimatter asymmetry
Spyros Basilakos,1,2 Nick E. Mavromatos ,3 and Joan Solà Peracaula 4
1
Academy of Athens, Research Center for Astronomy and Applied Mathematics,
Soranou Efessiou 4, 115 27 Athens, Greece
2
National Observatory of Athens, Lofos Nymfon, 11852 Athens, Greece
3
Theoretical Particle Physics and Cosmology Group, Physics Department,
King’s College London, Strand, London WC2R 2LS, United Kingdom
4
Departament de Física Quàntica i Astrofísica, and Institute of Cosmos Sciences (ICCUB),
Universitat de Barcelona, Avinguda Diagonal 647 E-08028 Barcelona, Catalonia, Spain
(Received 22 July 2019; accepted 8 January 2020; published 5 February 2020)
We present a model for the Universe in which quantum anomalies are argued to play an important dual
role: they are responsible for generating matter-antimatter asymmetry in the cosmos, but also provide time-
dependent contributions to the vacuum energy density of “running-vacuum” type, which drive the
Universe’s evolution. According to this scenario, during the inflationary phase of a string-inspired
Universe, and its subsequent exit, the existence of primordial gravitational waves induces gravitational
anomalies, which couple to the [Kalb-Ramond (KR)] axion field emerging from the antisymmetric tensor
field of the massless gravitational multiplet of the string. Such anomalous CP-violating interactions have
two important effects. First, they lead to contributions to the vacuum energy density of the form appearing
in the “running vacuum model” (RVM) framework, which are proportional to both, the square and the
fourth power of the effective Hubble parameter, H 2 and H4 respectively. The H4 terms may lead to
inflation, in a dynamical scenario whereby the role of the inflaton is played by the effective scalar-field
(“vacuumon”) representation of the RVM. Second, there is an undiluted KR axion at the end of inflation,
which plays an important role in generating matter-antimatter asymmetry in the cosmos, through
baryogenesis via leptogenesis in models involving heavy right-handed neutrinos. As the Universe exits
inflation and enters a radiation-dominated era, the generation of chiral fermionic matter is responsible for
the cancellation of gravitational anomalies, thus restoring diffeomorphism invariance for the matter/
radiation (quantum) theory, as required for consistency. Chiral U(1) anomalies may remain uncompensated,
though, during matter/radiation dominance, providing RVM-like H2 and H 4 contributions to the Universe
energy density. Finally, in the current era, when the Universe enters a de Sitter phase again, and matter is no
longer dominant, gravitational anomalies resurface, leading to RVM-like H2 contributions to the vacuum
energy density, which are however much more suppressed, as compared to their counterparts during
inflation, due to the smallness of the present era’s Hubble parameter H0. In turn, this feature endows the
observed dark energy with a dynamical character that follows the RVM pattern, a fact which has been
shown to improve the global fits to the current cosmological observations as compared to the concordance
ΛCDM model with its rigid cosmological constant, Λ > 0. Our model favors axionic dark matter, the
source of which can be the KR axion. The uncompensated chiral anomalies in late epochs of the Universe
are argued to play an important role in this, in the context of cosmological models characterized by the
presence of large-scale cosmic magnetic fields at late eras.
DOI: 10.1103/PhysRevD.101.045001
I. INTRODUCTION AND MOTIVATION: RUNNING

VACUUM MODEL FOR THE UNIVERSE
Over the last two decades, a plethora of cosmological
observations [1] have drastically changed our perception of
Published by the American Physical Society under the terms of the Universe. Strong evidence points towards the fact that
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to the energy budget of the cosmos in the current epoch
the author(s) and the published article’s title, journal citation, consists mostly (∼69%) of an unknown form of energy
and DOI. Funded by SCOAP3. [termed “dark energy” (DE)], whose equation of state is
2470-0010=2020=101(4)=045001(36) 045001-1 Published by the American Physical Society

BASILAKOS, MAVROMATOS, and SOLÀ PERACAULA PHYS. REV. D 101, 045001 (2020)
close to that of a cosmological constant, w ≃ −1. In however is time dependent, ρΛRVM ðtÞ ¼ ΛðtÞ=8πG.1 Let
addition, ∼26% consists of “dark matter” (DM), and thus us emphasize, however, that the time dependence of the
only about ∼5% of the Universe’s energy budget corre- vacuum energy density in the RVM is only through the
sponds to the known form of matter which we call baryonic Hubble rate (and its time derivatives), i.e., ρΛRVM ðtÞ ¼
matter. The dominance of the DE component results in the _
ρΛRVM ðHðtÞ; HðtÞ; …Þ, in contrast to the old phenomeno-
observed acceleration of the Universe at late eras, while its logical time-evolving models [24]. This feature is connected
equation of state w ≃ −1 points towards the fact that the
to the renormalization group (RG) in curved space-time, as
Universe enters again, for a second time (the first being
we shall see below. Ordinary matter and radiation are on top
during inflation), a de Sitter–type phase.
of it. In this picture, the total stress-energy tensor reads
Let us remark that most of the phenomenological
description of the cosmological data has been obtained
T μν ¼ −gμν κ 2 ΛðtÞ þ T m Λ
μν ¼ −gμν ρRVM ðtÞ þ T μν ; ð1Þ
m
in the context of the cosmological constant cold dark
matter (ΛCDM) model, the standard or “concordance”
where the superscript “m” refers generically here to matter
model of cosmology, which is characterized by a positive
(dust) and radiation contributions, where κ2 ¼ 8πG ¼ M−2
cosmological constant Λ and its associated vacuum Pl
is the (four-space-time-dimensional) gravitational constant,
energy density, ρΛ ¼ Λ=8πG (with G being Newton’s
gravitational constant). The latter plays the role of DE, where G ¼ M −2 P is Newton’s constant, M P ¼ 1.22 ×
and in fact it is the canonical DE candidate. In most 1019 GeV isptheffiffiffiffiffiffi four-dimensional Planck mass scale, and
18
cases the data is fitted to the spatially flat six-parameter MPl ¼ MP = 8π ¼ 2.43 × 10 GeV is the reduced Planck
canonical version of the ΛCDM model, the so-called mass (we work in units of ℏ ¼ c ¼ 1 throughout).
“base ΛCDM” [1]. The simplicity of the ΛCDM model, The total energy density ρtotal is therefore
however, may be to the detriment of its ability to provide
a better description of the cosmological observations as a ρtotal ¼ ρΛRVM þ ρdust þ ρradiation ; ð2Þ
whole. In fact, this could be at the root of the observed
discrepancies or “tensions” which are being persistently where we use the notation ρΛRVM to represent the RVM
observed in some observables, as we shall discuss contribution.
later on. The following renormalization group equation (RGE)
An important question is therefore whether the current de was proposed for ρΛRVM in the context of the RVM as a
Sitter phase of the Universe is due to the dominance of a function of the Hubble rate [3–6]:
purely cosmological-constant type DE, with w ¼ −1
exactly, or if there is a time-dependent vacuum energy dρΛRVM ðtÞ 1 X 2 2 4 H6
¼ ai Mi H þbi H þci 2 þ : ð3Þ
density that resembles to a good approximation the de Sitter dlnH2 ð4πÞ2 i Mi
phase. At a more fundamental level, the vacuum energy is
probably the result of quantum gravity effects, and in this Here H plays the role of the running scale μ of the RGE.
sense, understanding its microscopic nature might have to The coefficients ai ; bi ; ci … are dimensionless and they
wait for some time, until a satisfactory theory of quantum receive contributions from loop corrections of fermion
gravity, supported by observations, becomes available. This (i ¼ F) and boson (i ¼ B) matter fields with different
will also lead to a resolution of the longstanding cosmo- masses Mi . The missing term proportional to M 4i on the rhs
logical constant problem [2]. Nonetheless, like with all of the above equation is forbidden since there is no fully
other fundamental interactions in nature, there might be an active particle for the RGE, as all masses are larger than the
effective field theory description that captures the essential typical value of H at any epoch of the cosmological
features and is in agreement with observations, even evolution below the Planck mass. Therefore, the running
providing further insights for them. Such an attempt has goes slowly thanks to the decoupling terms. However,
been made by the development of the so-called “running
vacuum model” (RVM) [3–5] (see also Refs. [6,7] and 1
We note at this point that such a model for the Universe’s
references therein for a detailed review). Numerous studies vacuum energy has also been advocated within the context of
of that model’s cosmological evolution from the early string/brane universes in the presence of space-time brane defects
Universe to the present day can be found in Refs. [8–13]. [21]; quantum fluctuations of the latter induce a noncriticality of
the string universe [22], manifested through the generation of a
Furthermore, detailed confrontations with the recent cos- target-space vacuum energy dependence on the Liouville mode,
mological data has been presented in Refs. [14–16], which which is identified with the cosmic time [23]. Given the
extend the analyses of Refs. [17–19] and of older connection of the Liouville mode with a world-sheet local RG
works [20]. scale, this picture provides an interpretation of the cosmic time as
some sort of RG scale. Such a RG-like picture also lies at the
An important feature of the RVM is the existence of a heart of the RVM evolution, but from a rather different perspec-
“de Sitter–like” vacuum energy term in the total stress tive [3–5], not associated with specific string models, as we shall
tensor, with an equation of state wRVM ¼ −1, which review below.
045001-2
GRAVITATIONAL AND CHIRAL ANOMALIES IN THE RUNNING … PHYS. REV. D 101, 045001 (2020)
because of the dimensionality of ρΛ , the first allowed term _ ¼ a0 þ a1 H

ρΛRVM ðH; HÞ _ þ a2 H 2 þ a3 H
_ 2 þ a4 H 4
is a “soft decoupling term” ∼M 2i H 2 [3], which increases _ 2 þ
þ a5 HH ð4Þ
with the value of the masses and hence the effect need not
be negligible.2 For this reason the running is actually
where the coefficients ai have different dimensionalities in
dominated by the heaviest fields in the particular grand
natural units, and … denotes the possible decoupling terms
unified theory (GUT) context where the considerations are
(suppressed by mass powers) which are irrelevant for our
made [25]. This is in contrast to what happens in the usual
discussion. Specifically, a0 has dimension four since this is
gauge theories, like QED or QCD, where the decoupling
the dimension of ρΛ ; a1 and a2 have dimension two; and,
terms are all suppressed. The next-to-leading terms ∼H4
finally, a3 , a4 and a5 are dimensionless. The RVM is the
are not suppressed by heavy masses, and although irrel-
extension of the ΛCDM model based on a dynamical
evant for the current Universe, they can nonetheless play a
vacuum energy density of the form (4), stemming from the
central role in the early Universe and can explain inflation
basic RG equation (3). Despite the fact that higher-order
[7–10,26]. The conventional terms suppressed à la
terms are still possible in Eq. (4), the expression as written
Appelquist and Carazzone [27] appear only at the next-to-
contains the basic terms up to four derivatives of the scale
next-to-leading order, i.e., the OðH6 =M2i Þ terms of the cos-
factor, and hence it encodes the basic ingredients of the
mological RGE [6,28], which are a factor of H4 =M4i ≪ 1 model both for the low- (i.e., the late) and the high-energy
smaller than the soft decoupling (leading) ones. See (early and very early) Universe. In particular it encodes a
Ref. [3] for the original proposal and Ref. [29] for addi- possible description of inflation.
tional discussions. For simplicity, let us hereafter stick to the simplest
It is important to note that, because of the general association μ ¼ H. Taking into account that H _ ¼
covariance of the effective action, among the possible 2
−ðq þ 1ÞH , where q is the deceleration parameter,
terms emerging from the quantum effects one expects only which assumes the values q ¼ 1; 1=2; −1, as we move
those carrying an even number of time derivatives of the from the radiation- into the matter- and DE-dominated
scale factor a. If expressed in terms of the Hubble rate, epochs, respectively, we can see that the modification
_
H ¼ a=a, this amounts to terms of the form H2 , H,_ H4, H_ 2, _ is not very important and we can pick
introduced by H
H2 H_ etc. Thus, the linear terms in H (and in general
up the main effect already with the canonical association
any term with an odd number of derivatives of the scale μ ¼ H; this is indeed substantiated in the practical
factor, such as H3 , HH,_ ̈ etc.) are forbidden in the
H analyses (see e.g., Refs. [14–17]). In this situation, we
RVM since they would be incompatible with the general have a1 ¼ a3 ¼ a5 ¼ 0 in Eq. (4). The remaining coef-
covariance of the effective action [6]. In particular, at ficients can be related immediately to those in Eq. (3),
low energies only the H 2 and H _ terms are relevant for and the final result can be cast as [6]
the phenomenological confrontation with the data. The
higher-order ones can however be important for the early ΛðHÞ 3 H4
ρΛRVM ðHÞ 2
¼ 2 ¼ 2 c0 þ νH þ α 2 ; ð5Þ
Universe [8–13,26]. κ κ HI
As indicated, in Eq. (3) we have identified the RG scale μ
as μ ∼ H, and hence the Hubble rate plays the role of the where HI is the Hubble parameter close to the GUT
typical RG scale in cosmology. However, a more general scale, c0 is an integration constant (with mass dimension
option would be to associate μ2 to a linear combination of þ2 in natural units, i.e., energy squared), while the
H2 and H _ (both terms being dimensionally homogeneous). coefficients ðν; αÞ are written as [6,9]
Adopting this setting and integrating Eq. (3) up to the terms
of OðH4 Þ, or similar dimension, it is easy to see that we can 1 X M 2i
ν¼ ai ð6Þ
express the result as follows: 48π 2 i¼F;B MPl
and
2
We note that, even if we consider the radiation-dominated
1 H2I X
epoch of the Universe, at temperature T, the Friedmann equation α¼ bi : ð7Þ
(with ρm ∼ T 4 ) implies that the requirement of satisfying the 96π 2 M 2Pl i¼F;B
condition H > M i roughly means T 2 =M Pl > M i , or equivalently
M 4i =T 4 < M 2i =M 2Pl ≪ 1 for any particle of mass M i . Hence, at the In fact ν and α can be viewed as the reduced (dimen-
time when the ∼M4i contributions to the running of the vacuum sionless) beta functions of ρΛRVM at low and high energies
energy density start to become active, they are still negligible respectively [4–6]. Of course, due to the fact that all
compared to the radiation contribution ∼T 4 . We therefore
conclude that, within the RG formulation in which the RVM known particles have M2i ≪ M 2Pl , the above coefficients
is contextualized, the terms ∼M4i remain RG decoupled through- are expected to be quite small in a typical GUT, namely
out the entire cosmic history [6]. Oð10−6 –10−3 Þ; see Ref. [4].
045001-3
On considering a spatially flat Friedman-Lemaître- On the other hand, in the late universe, when the term
Robertson-Walker (FLRW) space-time, favored by obser- c0 =H2 in Eq. (8) begins to dominate over αH2 =H2I , the
vations [1], which we restrict our attention to in this work, corresponding integration leads to the solution
one can show that the main cosmological equations in the
presence of the RVM vacuum energy density (5) acquire H 2 ðaÞ ¼ H20 ½Ω̃m0 a−3ð1−νÞ þ Ω̃Λ0 ; ð11Þ
the form [8,9]
where Ω̃m0 ¼ Ω1−νm0
and Ω̃Λ0 ¼ 1 − Ω̃m0 ¼ Ω1−ν
Λ0 −ν
, with Ωm0 þ
_ 3 2 c0 H2 ΩΛ0 ¼ 1 being the standard sum rule (with the subscript
H þ ð1 þ ωÞH 1 − ν − 2 − α 2 ¼ 0; ð8Þ
2 H HI “0” denoting present-era quantities). The presence of the
parameter ν in the scaling of the matter contribution in
where ω ¼ ρm =pm , with ρm (pm ) being the matter/radiation Eq. (11) is an important and characteristic prediction of the
energy density (pressure), and the overdot denotes a RVM that allows comparison with the data.
derivative with respect to the cosmic time t of the In fact, the RVM agrees excellently with the current
FLRW universe. In the early universe we have relativistic cosmological data at large scales [1], but also makes
matter, ρm ¼ ρradiation with ω ¼ 1=3, while in the late important predictions [14–17] that could alleviate current
universe matter is dominated by dust, ρm ¼ ρdust with tensions in the data, concerning, for instance, the so-called
ω ¼ 0. Unlike the standard ΛCDM model of cosmology σ 8 tension and an associated improvement in describing
(Λ ¼ const) [1], here there is an exchange between matter large-scale structure formation, compared to the ΛCDM
and vacuum, which implies paradigm. The model also provides better insight into
the discrepancy with the (local) value of H0 between
ρ_ m þ 3ð1 þ ωÞHρm ¼ −_ρΛRVM : ð9Þ measurements by the Hubble Space Telescope, based
on Cepheid observations [30], and those by the Planck
The global dynamics of the RVM throughout the cosmic Collaboration, based on CMB studies [1]. In Refs. [14–16]
history has been studied in detail in Refs. [8,9]. According it was argued that the presence of the index ν in the RVM
to it, the universe starts from a nonsingular state charac- evolution of the Hubble parameter (11), which affects the
terized by an unstable initial de Sitter vacuum phase [11]. It scaling of the vacuum energy density (5) and, thus,
subsequently passes smoothly from an early inflationary differentiates it from the standard ΛCDM case, leads to
epoch to a radiation period (“graceful exit”) and, at the end, combined fits to SnIA þ BAO þ HðzÞ þ CMB data that
it goes into the dark-matter- and dark-energy-dominated favor a lower value of σ 8 .
epochs. The RVM evolution also provides an explanation Depending on whether one considers an interaction of
of the large entropy problem [7,11–13]. Below we briefly the dynamical DE with matter or assumes self-conservation
present the main points, for concreteness. Focusing on the of the DE, one can favor the lower value of H0 measured by
early universe era, for which c0 =H 2 ≪ 1, the integrated the Planck Collaboration [1] or push this value higher. This
form of Eq. (8) admits the following solution in terms of the feature has been demonstrated recently in Ref. [31], where
scale factor (upon using d=dt ¼ Hd=da in it): it was shown that, upon the assumptions that the DE adopts
the RVM form, and does not interact with matter, it is

1 − ν 1=2 HI possible to simultaneously decrease the value of σ 8 and
HðaÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð10Þ increase the prediction on H0 , such that the fitted value of
α Da3ð1−νÞð1þωm Þ þ 1
H0 definitely becomes much closer to the local value
where D > 0 is a constant. It is easy to check that for determined by Riess et al. [30].
Remarkably, some microscopic models supporting the
Da4ð1−νÞ ≪ 1 the universe starts from an unstable de Sitter
RVM-type evolution (5)–(11) of the energy density of
era H2 ¼ ð1 − νÞH2I =α which is powered by the huge value
pffiffiffi the Universe have been presented in Ref. [26], based
of HI ∼ αM 2X =MPl ≲ ð10−5 –10−6 ÞM Pl [7], where MX ∼ on inflationary scenarios involving dynamical breaking
1016 GeV is the typical value of the GUT scale. Note that of minimal supergravity, or in Refs. [5,6] on the basis of the
the previous relation is essential, since it is equivalent to the conformal anomaly-induced effective action. One of the
condition that the fluctuations from the tensor modes do not main points of the current work is to demonstrate that RVM
induce cosmic microwave background (CMB) temperature contributions of H2 type in the vacuum energy density arise
anisotropies larger than the observed ones (H=MPl ≲ 10−5 in more generic cosmological scenarios, inspired by string
in the early universe), and it is indeed satisfied for theory, in which axion fields, coupled to gravitational
α ∼ 10−3 –10−4 , which is in the expected range for this anomalies in de Sitter eras of the Universe, also result in
small parameter. After the early inflationary epoch, spe- RVM H2 contributions.
cifically in the case of Da4ð1−νÞ ≫ 1, we find H2 ∼ However, our work will make an important further
a3ð1−νÞð1þωm Þ ∼ a−4 (for jνj ≪ 1, ω ¼ 1=3) and the universe step by presenting a consistent (albeit minimal, rather
definitely enters the standard radiation phase, as expected. toy) scenario, of a string Universe, in which primordial
045001-4
gravitational waves induce gravitational anomalies during chiral anomaly, a large-scale cosmic magnetic field at late
the inflationary phase, where only the inflaton and gravi- epochs, whose magnetic energy density contributes to the
tational degrees of freedom (d.o.f.), including the Kalb- late-era energy budget of the Universe, with terms of RVM
Ramond (KR) axion associated with the antisymmetric type, scaling as H20 . There are models [43], however, in which
tensor field of the massless gravitational string multiplet, the KR axion couples to chiral matter (such as Majorana
are present in the string low-energy effective action [32– right-handed neutrinos) via shift-symmetry-breaking inter-
34]. The coupling of the KR axion to the gravitational actions, possibly generated by nonperturbative effects (string
anomaly leads to undiluted KR background fields at the instantons), and via shift-symmetry-preserving kinetic mix-
end of inflation, which spontaneously violate Lorentz ing to other axions that are abundant in string theory [44]. In
and CPT symmetry. This, in turn, plays an important role fact, such a mixing allows for the generation of a Majorana
in generating lepton asymmetry in models involving mass for the right-handed neutrinos, which is a crucial feature
(heavy) right-handed neutrinos [35–38], through the decays for the aforementioned leptogenesis scenario [36–38]. These
of the latter into Standard Model (SM) particles in the string theory axions can then play the role of additional
presence of the KR background. The lepton asymmetry can components of DM (in some of these scenarios, there is also a
then be communicated to the baryon sector via standard nonperturbative generated potential for the KR axion itself, at
baryon- (B) and lepton- (L) number-violating, but B − L- late eras, which thus implies its potential role as a massive
conserving, sphaleron processes in the SM sector of the DM candidate).
model [39]. In the current era, where matter becomes subdominant,
The basic results of this approach have already appeared and the Universe enters a de Sitter phase again, dominated
in Ref. [40]. Here we discuss the details but also present by dark energy, gravitational anomalies due to gravita-
further developments, in particular concerning the potential tional-wave perturbations resurface, but they are much
role of KR axions as dark matter in late eras of the more suppressed compared to their primordial counterparts,
Universe. since the current Hubble parameter H 0 is much more
Gravitational anomalies, when present, are known to suppressed compared to the one during inflation, HI ≫ H 0 .
affect the diffeomorphism invariance of the quantum As a matter of fact, this is also what makes it possible for
theory, in the sense that the matter stress-energy tensor the DE in our epoch to inherit a “relic” dynamical H 2
is not conserved [41]. In the absence of matter/radiation component as part of the observed DE contribution to the
d.o.f., as is the case in our string effective model during current energy budget of the cosmos. Therefore, in the
inflation, where we assume only d.o.f. from the gravita- context of the scenario described in the present article, we
tional string multiplet to be present, this may not be a naturally predict dynamical DE, which, as argued above,
catastrophe. The anomaly-induced nonconservation of the seems to be favored by current observations [14–17,45].
stress tensor simply accounts for the exchange of energy In the above scenario, therefore, the matter dominance
among the (quantum) gravitational d.o.f. over antimatter is entirely attributed to the existence of
During the radiation and matter eras, however, gravita- anomalies and the associated coupling of a gravitational
tional anomalies should be canceled for the consistency of axion d.o.f. (the KR axion) to them. In this work we shall
the matter quantum theory, which should be diffeomor- discuss all such issues in detail, with the aim of demon-
phism invariant. In our model this is provided by the strating the potential importance of gravitational anomalies
generation of chiral fermion matter with anomalous for the dominance of matter over antimatter in the cosmos
axial currents, such as chiral leptons in the SM sector and thus for our “very existence.” The H2 -RVM-type
or other chiral fermions that might exist in beyond the vacuum energy, associated with the anomaly contributions,
Standard Model physics models, which cancel the gravi- plus the existence of (“stiff”) axion DM, might then
tational anomalies during this epoch. The coupling of constitute smoking-gun evidence for such claims.
the (undiluted) KR axion to right-handed massive neutrino The structure of the article is as follows. In Sec. II A, we
matter during the early radiation era succeeding inflation discuss the (four-space-time-dimensional) primordial effec-
is essential for leptogenesis via the mechanism of tive action of the model, based only on the gravitational
Refs. [35–38]. d.o.f. of the massless bosonic string multiplet. By imposing
In general, chiral anomalies survive in the radiation- and the constraint on the modification of the Bianchi identity
matter-dominated eras, and this is crucial for providing a due to the gravitational Chern-Simons (gCS) terms by
link between the KR axion and the DM content of the means of a pseudoscalar Lagrange multiplier field in the
Universe at late epochs. The reader should recall that chiral path integral, we demonstrate how the latter acquires
anomalies are harmless from a diffeomorphism-invariance dynamics and becomes equivalent to a fully fledged KR
point of view, as they do not contribute to the stress tensor axion field. Its CP-violating coupling to the anomaly term
of matter. As we shall discuss in this article, the KR axion is crucial in ensuring background solutions, which sponta-
provides a source of (“stiff” [42]) axionic DM, and is neously break Lorentz and CPT symmetry, and remain
responsible for generating, through its coupling with the undiluted at the end of the inflationary era. This is
045001-5
demonstrated in Sec. II B, where it is also shown that spin-0 scalar field, the dilaton Φ,3 where gs ¼ eΦ is the
primordial gravitational waves are the primary source of string coupling, and the dimensionless spin-1 antisymmetric
gravitational anomalies during that phase in the Universe’s tensor (Kalb-Ramond) field Bμν ¼ −Bνμ . In the closed-
evolution. Moreover, the anomaly contributes to the energy string sector, to which we restrict ourselves for concreteness
density of the cosmic fluid terms which have the form of for the purposes of this work, there is a Uð1Þ gauge
RVM contributions, proportional to the square of the symmetry Bμν → Bμν þ ∂ μ θν − ∂ ν θμ which characterizes
Hubble parameter, H2 ðtÞ. In Sec. II C we discuss the the target-space low-energy string effective action. This
potential role of the gravitational anomaly term, averaged implies that the latter depends only on the field strength of
over the inflationary space-time, as a provider of an the field Bμν , which is a three-form with components
effective H4 term in the RVM energy density, which can
then be held responsible for inflation, without the need for Hμνρ ¼ ∂ ½μ Bνρ ; ð12Þ
invoking an external inflaton field, the role of which is thus
played by the scalar-field (the “vacuumon”) effective where the symbol ½… denotes complete antisymmetrization
description of the RVM. In Sec. III A, we discuss the of the respective indices. The three-form Hμνρ satisfies the
cancellation of the gravitational anomalies during radia- Bianchi identity
tion-/matter-dominated eras, as a result of the generation of
anomalous chiral leptonic matter at the end of inflation. ∂ ½μ Hνρσ ¼ 0; ð13Þ
There remain, however, uncompensated chiral anomalies
during those eras, which also furnish the cosmic-fluid by construction.
energy density with RVM-like H 2 ðtÞ contributions. The The bosonic part of the (four-space-time-dimensional)
presence of the (undiluted by inflation) KR axion field, effective action, SB , that reproduces the string scattering
plays an important role in generating leptogenesis during amplitudes to lowest nontrivial order in an expansion in
the radiation era (and subsequently baryogenesis) in powers of the string Regge slope α0 (i.e., quadratic order in
models with heavy right-handed sterile neutrinos, which derivatives), to which we restrict our attention from now on,
is discussed in Sec. III B. In Sec. IV, we demonstrate how reads in the Einstein frame [33,34]4
the KR axion background, which couples to the uncom-
Z
pensated chiral anomaly, plays a role analogous to the 4 pffiffiffiffiffiffi 1 1
chiral chemical potential in the electrodynamics of standard SB ¼ d x −g 2
½−R þ 2∂ μ Φ∂ μ Φ − e−4Φ Hλμν Hλμν
2κ 6
axions, which has important implications for the genera-
2
tion of a cosmological magnetic field at late eras of the − 0 2 e2Φ δc þ ; ð14Þ
Universe, whose energy density contributes to the axion- 3α κ
DM energy budget. In Sec. V we speculate on extensions of
the model, involving mixing of the KR axion with other where Hμνρ ≡ κ −1 Hμνρ has dimensions of ½mass2 , and the
axions, which exist abundantly in string theory [44] and … represents higher-derivative terms, which are of higher
can play the role of additional axionic DM components. order in α0 , where α0 ¼ M −2
s is the Regge slope of the string
We discuss the compatibility of the generation of a shift- and Ms is the string mass scale. The latter is not necessarily
symmetry-breaking quintessence-like potential for the KR
field at late eras of the Universe with the (approximately) 3
The dilaton is sometimes referred to as the trace part of the
constant background configurations that we studied in graviton. This has the following meaning. If we apply the
Sec. IV. Finally, Sec. VI contains our conclusions. equivalence principle, so that locally the target space-time, in
Although in this work we consider the concrete case in which a string propagates, is taken—through an appropriate
coordinate choice—to be the flat Minkowski space-time, then
which the string mass scale is of the order of the (reduced) the graviton fluctuations are defined through the linearization of
Planck mass, our results are valid in the more general case the metric tensor: gμν ¼ ημν þ κhμν , where hμν is a mass-dimen-
where these scales are different. A brief discussion on this sion-one tensor with respect to the Lorentz symmetry, and κ 2 ¼
is given in the Appendix. 8πG is the four-dimensional gravitational constant. The associated
group SO(D − 1, 1) of transformations in D target-space dimen-
sions of the string contains then a traceless spin-2 tensor repre-
II. ANOMALOUS STRING EFFECTIVE ACTIONS, sentation, corresponding to the graviton, the spin-1 antisymmetric
INFLATION AND RUNNING VACUUM part, and a trace part, which refers to as the dilaton κ −1 Φ, with Φ
dimensionless. In general relativity, one imposes a “gauge fixing,”
A. The primordial effective action with in which the graviton fluctuation tensor in the linearized formalism
(gravitational) anomalies is transverse and traceless, thus corresponding to the aforemen-
tioned spin-2 traceless part of the SOðD − 1; 1Þ representations.
The massless bosonic gravitational multiplet of a generic 4
The conventions and definitions we use throughout this work
string theory consists of three fields [32]: a traceless, are as follows: the metric signature ðþ; −; −; −Þ, Riemann
symmetric, dimensionless, spin-2 tensor field gμν , that is curvature tensor Rλ μνσ ¼∂ ν Γλ μσ þΓρ μσ Γλ ρν −ðν↔σÞ, Ricci tensor
uniquely identified with the graviton, a dimensionless Rμν ¼ Rλ μλν , and Ricci scalar R ¼ Rμν gμν.
045001-6
the same as the four-dimensional gravitational con- where we used differential-form language for brevity, with
stant κ 2 ¼ 8πG ¼ M −2 Pl . ∧ denoting the usual exterior (“wedge”) product among
The last term on the right-hand side of Eq. (14) repre- differential forms, such that fðkÞ ∧ gðlÞ ¼ ð−1Þkl gðlÞ ∧ fðkÞ ,
sents a (four-space-time-dimensional) vacuum energy term. where fðkÞ , and gðlÞ are k− and l− forms, respectively.
In noncritical string models [22], such a term arises from a Above, A is the Yang-Mills potential (gauge field) one-
positive δc > 0 central charge surplus of supercritical form, and ωa b is the spin connection one-form [the latin
strings, which owes its existence to σ-model conformal indices a, b, c, d are tangent space i.e., Lorentz group
anomaly contributions from “internal dimensions” of the SO(1,3) indices]. The addition of Eq. (17) leads to a
string, the “external dimensions” D ¼ 4 defining the four- modification of the Bianchi identity (13) [32]
dimensional target space-time of our Universe. In brane-
universe scenarios, such vacuum energy contributions could α0
come from bulk-space terms, and they include anti–de dH ¼ TrðR ∧ R − F ∧ FÞ ð18Þ
8κ
Sitter–type (negative) contributions [46]. For our purposes
in this work we shall assume δc ¼ 0. We shall also assume where F ¼ dA þ A ∧ A is the Yang-Mills field strength
that the dilaton varies slowly or that it has stabilized (through two-form and Ra b ¼ dωa b þ ωa c ∧ ωc b is the curvature
some appropriate nonperturbative string mechanism) to a two-form and the trace (Tr) is over gauge and Lorentz
constant value Φ0 , so that we may approximate ∂ μ Φ∂ μ Φ ≃ 0 group indices. The nonzero quantity on the right-hand side
in Eq. (14) throughout the current work. This implies an of Eq. (18) is the “mixed (gauge and gravitational) quantum
(approximately) constant string coupling gs ¼ gs eΦ0 .
ð0Þ anomaly”.5
Without loss of generality, then, we may set Φ0 ¼ 0. The Bianchi identity constraint (18) in differential-form
ð0Þ language can be expressed in the usual tensor notation as
The string coupling gs can be fixed by phenomenological follows:
considerations of the four-dimensional effective field
theory [32]. α0 pffiffiffiffiffiffi
εabc μ Habc ;μ ¼ −gðRμνρσ R̃μνρσ − Fμν F̃μν Þ
We can then write the action SB as 32κ
Z pffiffiffiffiffiffi
pffiffiffiffiffiffi 1 1 ≡ −gGðω; AÞ; ð19Þ
SB ¼ − d4 x −g R þ H H λμν
þ : ð15Þ
2κ 2 6 λμν where the semicolon denotes a covariant derivative with
It is known [32,33] that the KR field strength terms H in 2 respect to the standard Christoffel connection, and
Eq. (15) can be absorbed (up to an irrelevant total
pffiffiffiffiffiffi sgnðgÞ
divergence) into a contorted generalized curvature R̄ðΓ̄Þ, εμνρσ ¼ −gϵμνρσ ; εμνρσ ¼ pffiffiffiffiffiffi ϵμνρσ ; ð20Þ
−g
with a “torsional connection” [47] Γ̄, corresponding to a
contorsion tensor proportional to the field strength Hρμν ,
where ϵ0123 ¼ þ1, etc., are the gravitationally covariant
κ Levi-Civita tensor densities, totally antisymmetric in their
Γ̄ρμν ¼ Γρμν þ pffiffiffi Hρμν ≠ Γ̄ρνμ ; ð16Þ g over the curvature or gauge field-
3 indices. The symbol ð…Þ
strength tensors denotes the corresponding dual, defined as
where Γρμν ¼ Γρνμ is the torsion-free Christoffel symbol.
Exploiting local field redefinition ambiguities [33,34], 1 1
R̃μνρσ ¼ εμνλπ Rλπ ρσ ; F̃μν ¼ εμνρσ Fρσ : ð21Þ
which do not affect the perturbative scattering amplitudes, 2 2
one may extend the above conclusion to the quartic order
in derivatives, that is, to the Oðα0 Þ effective low-energy Since the anomaly Gðω; AÞ is an exact one-loop result,
action, which includes Gauss-Bonnet quadratic curvature one may implement the Bianchi identity (19) as a δ-
invariants. functional constraint in the quantum path integral of the
In string theory, in the presence of gauge and gravita- action (15) over the fields H, A, and gμν , and express the
tional fields, the cancellation of anomalies, requires the latter in terms
pffiffiffiof a Lagrange multiplier (pseudoscalar)
pfield
ffiffiffi
modification of the right-hand side of Eq. (12) by appro- [34] bðxÞ= 3 [where the normalization factor 3 is
priate gauge [Yang-Mills (Y)] and Lorentz (L) Chern- inserted so that the field bðxÞ will acquire a canonical
Simons three-forms [32] kinetic term, as we shall see below]:
α0
H ¼ dB þ ðΩ − Ω3Y Þ; 5
8κ 3L
Notice that the modifications (17) and the right-hand side of
the Bianchi identity (18) contain the torsion-free spin connection.
2 In fact, it can be shown [48,49] that any potential contributions
Ω3L ¼ ωa c ∧ dωc a þ ωa c ∧ ωc d ∧ ωd a ;
3 from the torsion H three-form in the anomaly equation can be
removed by adding to the string effective action appropriate
Ω3Y ¼ A ∧ dA þ A ∧ A ∧ A; ð17Þ counterterms order by order in perturbation theory.
045001-7
Z Z
μνρσ 1
4 pffiffiffiffiffiffi p μνρσ
Πx δðε Hνρσ ðxÞ;μ − Gðω; AÞÞ ⇒ Db exp i d x −g ffiffiffi bðxÞðε Hνρσ ðxÞ;μ − Gðω; AÞÞ
3
Z Z
4 pffiffiffiffiffiffi μ 1 νρσ bðxÞ
¼ Db exp −i d x −g ∂ bðxÞ p ffiffi
ffi ϵμνρσ H þ p ffiffi
ffi Gðω; AÞ ð22Þ
3 3
where the second equality has been obtained by partial integration, upon assuming that the KR field strength dies
out at spatial infinity. Inserting Eq. (22) into the path integral with respect to the action (15), and integrating over the H
field, one obtains an effective action in terms of the anomaly and a canonically normalized dynamical, massless, KR axion
field bðxÞ [34]
Z rffiffiffi
4 pffiffiffiffiffiffi 1 1 μ 2 α0 μνρσ μν
Seff ¼ d x −g − 2 R þ ∂ μ b∂ b þ bðxÞðRμνρσ R̃ − Fμν F̃ Þ þ ; ð23Þ
B
2κ 2 3 96κ
where the dots … denote gauge, as well as higher- involving only the KR axion and the gravitational field. The
derivative, terms appearing in the string effective presence of the axion bðxÞ represents the effects of “torsion,”
action, which we ignore for our discussion here.6 We thus in view of our previous discussion on the role of the KR field
observe that, in view of the anomaly, the KR axion field strength as a (quantum) torsion [Eq. (16)] in string theory
couples to the gravitational and gauge fields. This inter- [32–34]. On ignoring the gauge sector, the topological
action is P and T violating, and hence in view of the overall density (24) becomes
CPT invariance of the quantum theory, also CP violating. It
will be quite important for our purposes in this work. In pffiffiffiffiffiffi pffiffiffiffiffiffi pffiffiffiffiffiffi
pffiffiffiffiffiffi −gRμνρσ R̃μνρσ ¼ −gKμ ðωÞ;μ ¼ ∂ μ ð −gKμ ðωÞÞ
fact, the term −gðRμνρσ R̃μνρσ − Fμν F̃μν Þ in Eq. (23) is the
well-known Hirzebruch-Pontryagin topological density μναβ ab 2
¼ 2∂ μ ϵ ων ∂ α ωβab þ ωαa ωβcb ;
c
and is a total derivative 3
pffiffiffiffiffiffi ð26Þ
−gðRμνρσ R̃μνρσ − Fμν F̃μν Þ
pffiffiffiffiffiffi pffiffiffiffiffiffi
¼ −gKμmixed ðωÞ;μ ¼ ∂ μ ð −gKμmixed ðωÞÞ which is also called the “gravitational Chern-Simons” term, a
terminology that we shall use in this work. The (purely
2 c pffiffiffiffiffiffi
¼ 2∂ μ ϵμναβ ωab ν ∂ ω
α βab þ ω ω gravitational) quantity −gKμ may be viewed as the “axial
3 αa βcb
current density” of our bosonic theory (i.e., in the absence of
μναβ 2 ijk i j k fermions), as its (covariant) four-divergence is related to the
− 2ϵ Aν ∂ α Aβ þ f Aν Aα Aβ ;
i i
ð24Þ
3 gravitational anomaly. For completeness and future conven-
pffiffiffiffiffiffi
ience, we also express below −gKμ in terms of the
with latin letters i, j, k being gauge group indices, and
pffiffiffiffiffiffi μ (standard) torsion-free Christoffel connection Γα βγ ,
−gKmixed denoting the mixed (gauge and gravitational)
anomaly current density.
pffiffiffiffiffiffi μ μβγδ ν σ 2 ν σ λ
In the early Universe, before and during inflation, we −gK ¼ ϵ Γ βσ ∂ γ Γ δν þ Γ βσ Γ γλ Γ δν : ð27Þ
assume that only fields from the gravitational multiplet of the 3
string exist, which implies that our effective action pertinent
to the dynamics of the inflationary period, is given by Eq. (23) We now notice that, by partially integrating the CP-
upon setting the gauge fields to zero, A ¼ 0. Thus, to describe violating anomaly term in Eq. (25), ignoring surface terms
the dynamics of the beginning and the inflationary period of (on account of the assumption that the gravitational field
the Universe, we use the effective action and its derivatives vanish at infinity), and using Eq. (26),
one arrives at the effective action
Z
4 xpffiffiffiffiffiffi 1 1 Z
Seff ¼ d −g − 2 R þ ∂ μ b∂ μ b 4 pffiffiffiffiffiffi 1 1
B
2κ 2 Seff ¼ d x −g − 2 R þ ∂ μ b∂ μ b
rffiffiffi
B
2κ 2
2 α0 rffiffiffi
þ bðxÞRμνρσ R̃μνρσ þ ; ð25Þ 2 α0
3 96κ − ∂ μ bðxÞKμ þ
3 96κ
6 ≡ S grav þ S b þ S b-grav ; ð28Þ
It should be noticed that, in our conventions for the Levi-
Civita tensor (20), the kinetic term of the b field in Eq. (23) has
the opposite sign to that of the (covariant) square of the Hμνρ where S grav denotes the pure-gravity Einstein-Hilbert Ricci
tensor in Eq. (15). scalar action, Sb ðb; gαβ Þ denotes the “matter” action of the
045001-8
bðxÞ field, which does not contain derivatives of the we may write the corresponding (generic) Einstein equation
graviton, in the form
Z rffiffiffi
pffiffiffiffiffiffi 1 1 2 α0 κ μν
S ≡ d4 x −g ∂ μ b∂ μ b;
b ð29Þ Rμν − gμν R ¼ Λgμν þ C þ κ2 T μν ð36Þ
2 matter ;
2 3 12
and where T μνmatter is a generic matter stress tensor, including
rffiffiffi Z axion-like fields [like our KR axion above, cf. Eq. (29)]
2 α0 pffiffiffiffiffiffi which does not contain couplings to curvature and, in
S b-grav
≡− d4 x −gð∂ μ bðxÞKμ Þ
3 96κ general, derivatives of the metric tensor. The latter cou-
rffiffiffi Z
2 α0 pffiffiffiffiffiffi plings contribute only to Cμν . In standard situations, without
¼ d4 x −gbRμνρσ R̃μνρσ ; ð30Þ gravitational anomalies, general coordinate diffeomor-
3 96κ
phism invariance, implies the conservation of the matter
is the KR-axion-gravitational anomaly term (24). stress tensor, T μν
matter;ν ¼ 0, given the covariant constancy of
The “matter” KR-axion stress-energy tensor is calculated the metric, which ensures that the cosmological constant Λ
from Eq. (28) by using the standard definition of T bμν in contribution to the total energy-momentum tensor is con-
general relativity, served. Because of the curvature tensor Bianchi identity,
the Einstein tensor Rμν − 12 gμν R, also obeys such a covar-
2 δSb ðb; gαβ Þ 1 iant conservation law, but this is not the case for the Cotton
T bμν ¼ pffiffiffiffiffiffi μν ¼ ∂ μ b∂ ν b − gμν ð∂ α b∂ α bÞ: ð31Þ
−g δg 2 tensor, as one can readily check [41]:
To compute the metric variation of Eq. (30), we take into 1

Cμν ;μ ¼ − vν Rαβγδ R̃αβγδ : ð37Þ
account that the variation of the Christoffel symbol with 8
respect to the metric tensor gμν is
Thus, in the presence of gravitational anomalies, the
1 diffeomorphism invariance would appear to be in trouble,
δΓβ αγ ¼ gβδ ððδgδγ Þ;α þ ðδgαδ Þ;γ − ðδgαγ Þ;δ Þ: ð32Þ unless one deals with specific gravitational backgrounds
2
[41,50], such as the ones pertaining to the FLRW universe
One can then easily express the infinitesimal metric of interest to us here, for which the Pontryagin density
variation of the Pontryagin-term bRR̃ in terms of the so- vanishes Rμνρσ R̃μνρσ ¼ 0. Indeed, in our case, during the
called four-dimensional Cotton tensor Cμν [41]: inflationary era, for which A ¼ 0, the term bRR̃ in Eq. (25),
Z Z yields, on account of Eqs. (33)–(34), a Cotton tensor of the
pffiffiffiffiffiffi
4 μνρσ pffiffiffiffiffiffi form [34]
δ d x −gbRμνρσ R̃ ¼ 4 d4 x −gCμν δgμν
Z
pffiffiffiffiffiffi Cμν ∝ ð∂ ρ bR̃ρμλν Þ;λ þ ðμ ↔ νÞ; ð38Þ
¼ −4 d4 x −gCμν δgμν ; ð33Þ
where the dual Riemann tensor R̃μνρσ has been defined in
where [34,41] Eq. (21), and the numerical proportionality coefficients are
of no interest to us, and hence we do not write them
1 explicitly here. For a homogeneous and isotropic FLRW
C ≡ − vσ ðεσμαβ Rν β;α þ εσναβ Rμ β;α Þ þ vστ ðR̃τμσν þ R̃τνσμ Þ
μν
2 space-time, and axion field bðtÞ, for which only the
1 temporal derivative is nonzero, we obtain from Eq. (38)
¼ − ½ðvσ R̃λμσν Þ;λ þ ðμ ↔ νÞ;
00 ¼ 0, on account of the antisymmetry of the
that T bR R̃
2
vσ ≡ ∂ σ b ¼ b;σ ; vστ ≡ vτ;σ ¼ b;τ;σ : ð34Þ Riemann tensor Rμναβ ¼ −Rνμαβ and the properties of its
dual. The pressure density contributions of such terms also
As follows from its definition (34), and the properties of the vanish, as follows from the Bianchi identity of the Riemann
Riemann tensor, the Cotton tensor is traceless [41] curvature tensor, Rμ½νρσ ¼ 0, with ½… denoting antisym-
metrization of the respective indices. Thus for a FLRW
gμν Cμν ¼ 0: ð35Þ universe, the Cotton tensor vanishes, consistent with
diffeomorphism invariance.
At this stage, we would like to make some generic The “apparent” nonconservation of the matter stress
remarks concerning the conservation properties of the tensor in the presence of the Cotton tensor in the Einstein
Cotton tensor, and thus potential problems associated with equation (36) appears to be in contradiction with the
theories with gravitational anomalies [41]. From Eq. (33), perfectly covariant form of the axion-gCS coupling in
045001-9
Eq. (25) under general coordinate transformations. As is interaction terms in the effective action (28) are fully
standard, when evaluating anomalies in higher-order consistent, both formally and conceptually, as was to be
quantum-corrected effective actions, one employs specific expected given that such terms arise in the context of string
regularizations, such that there is some sort of conserved theory [34], which is a consistent theory of quantum
“improved” second-rank tensor which plays the role of gravity [32].
the energy-momentum tensor, compatible with general What we say here is that, rather than restricting [41,50]
covariance. the consistent space-time backgrounds by demanding the
This also happens here, for generic space-time back- covariant conservation of the pure axion matter stress
grounds. Indeed, as can be seen from Eq. (36), from the tensor, T μν
axion-matter;ν ¼ 0, when there are gravitational
Bianchi identities of the Einstein tensor, there is a con- anomalies, to which these axions couple, this on-shell
served modified stress-energy tensor conservation law breaks down, as a result of the exchange
rffiffiffi of energy with the gravitational field. There is instead a
2 α0 κ μν modified stress tensor (39) which remains conserved. This
κ 2 T̃ μν ≡ C þ κ 2 T μν
b þ Λg
μν
peculiarity refers only to axion fields and should be
bþΛþgCS
3 12
reflected in the solutions to the equations of motion for
⇒ T̃ μν
bþΛþgCS;μ ¼ 0; ð39Þ these fields in the presence of anomalous background
space-times. It is the purpose of this work to demonstrate
with the extra terms, proportional to the Cotton tensor Cμν, the existence of such consistent solutions; however as we
describing energy exchange between the axion and gravi- shall see they “spontaneously” violate Lorentz symmetry.
tational field.7 The covariant conservation law (39), then, The latter should not come as a surprise, due to the
leads as usual to the energy conservation of the KR axion- “spontaneous” breaking of diffeomorphism invariance by
gravity system. Any regularization scheme employed in the the anomalous space-time gravitational backgrounds (“vac-
computation of the anomaly should then respect this uum”). The underlying UV-complete, full quantum gravity
conservation law, and thus the CP-violating axion-gravity theory should be diffeomorphism invariant, as is the case in
string theory in our example.
7
The existence of a conserved modified stress-energy Indeed, as we shall discuss later on (cf. Secs. II B
tensor proportional to higher-curvature terms also characterizes and II C), primordial gravitational waves during the infla-
the case of dilaton-Gauss-Bonnet Oðα0 Þ terms in string effective tionary phase of FLRW universes do induce nontrivial
actions [51]:
CP-violating anomalous gravitational-Chern-Simons-KR-
Z axion couplings, and condensates of the gravitational
4pffiffiffiffiffiffi 1
SB ¼ d x −g ½−R þ 2∂ μ Φ∂ μ Φ anomaly. Upon taking into account such condensates,
2κ 2
the classical equation of motion for the KR axion field
α0
− c1 ð0Þ2
e−2Φ ðRμνρσ Rμνρσ − 4Rμν Rμν þ R2 Þ is modified from the standard one in the absence of
8gs κ 2 gravitational anomalies. In the anomaly-free case, the
þ c2 ð∂ΦÞ4 þ Þ ð40Þ KR axion, classically satisfies □bðxÞ ¼ 0, where □ is
the covariant D’ Alembertian. This implies the classical
where ci , i ¼ 1, 2 are numerical coefficients, and the … denotes conservation law T μν μν
b ;μ ¼ 0, where T b is given in Eq. (31).
contributions from the central charge deficit, antisymmetric
tensor fields, and higher derivatives. The existence of the As we shall see below, in the presence of gravitational-
Gauss-Bonnet terms, leads to a modified stress tensor conserva- anomaly condensates, this equation is modified to Eq. (45),
tion T ϕþGB;μ
μν ¼ 0, with which admits the nontrivial Lorentz-violating solutions
(72), mentioned above. For such solutions, it is the
2 1 α0 modified stress tensor (39), taking into account the KR-
T ϕþGB
μν ¼ ∂ Φ∂ ν Φ − 2 gμν ð∂ σ ΦÞ2 þ 2 P μν þ ;
2 μ
κ κ κ gravitational-Chern-Simons interaction, that is classically
c1 conserved on account of the classical Einstein equations.
P μν ¼ ð0Þ2 ðgμρ gνλ þ gνρ gμλ Þε ðR̃ργ αβ ∂ σ ½e−2Φ Þ;γ ;
σλαβ
ð41Þ
8gs As we shall discuss in Sec. II C, the anomaly condensates
induce a background FLRW universe with a positive (de
with the extra interactions describing energy exchange between Sitter–type) cosmological constant, which drives inflation.
dilatons Φ and gravity, in a similar spirit to our case with axions
interacting with the gravitational Chern-Simons term, which is However, in this case things are even more subtle, in the
also a higher-curvature term. In fact, the analogy with our case sense that the anomalous gravitational contributions are
goes even further, in the sense that the modifications P μν in obtained by averaging over quantum graviton fluctuations
Eq. (41) lead to nonpositive contributions to the modified stress- [52], and in this sense Einstein’s equations, which are
energy tensor, which in Ref. [51] were deemed important in classical equations, do not describe the graviton quantum
evading the no-hair theorem, allowing for black hole solutions
with (secondary) dilaton hair. In our case, the negative contri- fluctuations that induce the Chern-Simons term. Hence,
bution of the gravitational Chern-Simons term proves crucial for there is no inconsistency as far as the underlying quantum
the consistency of our framework; see Sec. II C. gravity (string) theory during the inflationary era of the
045001-10
Universe is concerned, which at low energies is described taking into account Eqs. (33), (34) and (37), the reader can
by the effective action (25), and contains only fields from easily verify that this equation implies the conservation of
the string gravitational multiplet. The latter is fully con- the improved stress tensor (39), as explained previously.
sistent and diffeomorphism invariant. In order to not disturb the homogeneity and isotropy of
Nonetheless, as we shall see in Sec. III A that, in the the inflationary space-time, we may assume only a (cosmic)
radiation or matter era, after the exit from inflation, the time t dependence of the KR background b̄ðtÞ, which, in
generation of chiral matter would lead to a cancellation of view of Eq. (45), would imply that only the temporal
gravitational anomalies, as would be “conventionally” component (μ ¼ 0) of the “axial current density” could be
required for the “consistency” of the matter and radiation nontrivial, K0 ðtÞ ≠ 0. The general solution of Eq. (45),
quantum field theory, without the need to employ a which we assume from now on, is
generalized stress-energy tensor [Eq. (39)]. In such a case, rffiffiffi
the axion fields would only couple at most to chiral or in _b̄ ¼ pCffiffiffiffiffiffi
0 2 α0 0
þ K; ð46Þ
general triangle anomalies, which do not contribute to the −g 3 96κ
stress tensor, due to their topological form [see the dis-
cussion in Sec. III A, after Eq. (99)], and, thus, the conven- where b̄_ ¼ dtd b̄ðtÞ and C0 is a constant. Equation (46) is a
tional local covariant conservation of the matter/radiation mathematically consistent relation, since both ∂ μ b and Kμ
stress tensor is guaranteed for any metric background. are (covariant) axial four-vectors.
After these important remarks we next proceed to discuss The relation (46) induces a background for the KR axion
the equation of state of the KR fluid. From Eq. (31), and field that spontaneously breaks Lorentz, CP and CPT
taking into account the generic relation for the stress-energy symmetry. In fact the masslessness of the KR axion b can
tensor for an observer moving with a four-velocity uμ with be understood by viewing this pseudoscalar field as the
respect to an inertial frame Goldstone boson of the spontaneously broken Lorentz
symmetry [22].
T μν ¼ ðρ þ pÞuμ uν − gμν p; ð42Þ The term proportional to C0 in Eq. (46) is expected to be
suppressed in an inflationary space-time, so without loss of
we obtain for the energy density ρb ¼ T brest
00 and pressure generality we may set from now on C0 ¼ 0 and consider the
p defined via T ii ¼ −p gii (no sum over i) for an
b brest b solution
observer at rest with respect to the cosmic frame of a FLRW rffiffiffi
2 α0 0
b̄_ ¼
Universe, with a homogeneous and isotropic KR axion
K: ð47Þ
field bðtÞ fluid, 3 96κ
1 1 From the anomaly equation (26), assuming
pffiffiffiffiffiffiffiffiffiffiffi homogeneity
ρb ¼ ðb̄_ Þ2 ; pb ¼ ðb̄_ Þ2 ¼ ρb : ð43Þ
2 2 and isotropy for the anomaly density −gðtÞKμ ðtÞ, where t
is the cosmic time, one has
This has a stiff matter [42] equation of state, w ¼ 1 and
hence cannot by itself lead to a “running vacuum” type of d pffiffiffiffiffiffi 0 pffiffiffiffiffiffi
fluid. The scaling (with the universe’s scale factor) of the ð −gK ðtÞÞ ¼ h −gRμνρσ R̃μνρσ i; ð48Þ
dt
energy density of stiff matter is
where h…i denotes appropriate averages over graviton
ρb ¼ pb ∼ a−3ð1þwÞ ¼ a−6 ; w ¼ 1: ð44Þ fluctuations in the inflationary space-time to be defined
below [52].
Below we shall explicitly demonstrate this by evaluating In an unperturbed FLRW space-time, with scale factor
the induced energy density, as a self-consistency check of aðtÞ, the right-hand side of Eq. (48) vanishes, as already
the approach. To this end, we first observe from Eq. (28), mentioned, which would imply
that the classical equations of motion of the KR axion field pffiffiffiffiffiffiffiffiffiffiffi
bðxÞ, imply the existence of backgrounds b̄ that satisfy K0 ðtÞ ∝ ð −gðtÞÞ−1 ∼ a−3 ðtÞ; ð49Þ
rffiffiffi
pffiffiffiffiffiffi α 2 α0 α consistent with the expected “stiff matter” scaling (44) in
∂ α −g ∂ b̄ − K ¼ 0; ð45Þ this case, where only a massless KR axion field without a
3 96κ
potential is the only constituent of “matter” in the Universe.
where, as we shall see, Kμ will be associated with an
average of the Hirzebruch-Pontryagin density (26) over the B. Gravitational waves during inflation, anomalies
inflationary space time, which in the presence of the CP- and a “running vacuum”
violating anomalous interactions of Eq. (28) can be non- In this context, another scalar field or mechanism, can be
vanishing [52]. By multiplying Eq. (45) with ∂ ν b, and introduced to induce inflation. At the moment we assume
045001-11
that the new field is some conventional inflaton field, φ, gravitational-wave perturbations [52]). This implies that
imported from an external framework which the current one the energy density ρφþb and pressure pφþb are
might be embedded into. Later on, in the next subsection,
1 1
we will see that such a scalar field need not be a new ρφþb ¼ ðb̄_ Þ2 þ ðφÞ
_ 2 þ UðφÞ;
fundamental field but just the one that enables mapping the 2 2
1 1
RVM to its scalar field representation. However, everything pφþb ¼ ðb̄_ Þ2 þ ðφÞ
_ 2 − UðφÞ: ð52Þ
that we will say below does not depend on the nature of φ 2 2
and hence we postpone the discussion of the scalar picture For slow running of both the φðtÞ and bðtÞ fields, that is
of the RVM to Sec. II C. So, let us assume for concreteness _ 2 ≪ jUðφÞj which we assume for our purposes
_ 2 , ðbÞ
ðφÞ
the existence of an inflaton scalar field, φ, which is different here (and we shall check the self-consistency of this
from the KR axion bðxÞ.8 Augmenting our effective action assumption explicitly in what follows), we observe then
(28) by the inclusion of a scalar-φ sector, with a canonical that the conditions for inflation are satisfied to leading order
kinetic term and a potential UðφÞ, we write for the complete in small quantities,
effective action9
Z pφþb ≃ −ρφþb ≃ −UðφÞ; ð53Þ
4 pffiffiffiffiffiffi 1 1
Seff ¼ d x −g − 2 R þ ∂ μ φ∂ μ φ − UðφÞ
bþφþgravity
2κ 2 provided UðφÞ > 0 (in our conventions).
rffiffiffi Naively speaking, as follows from Eq. (49), one would
1 2 α0
þ ∂ μ b∂ μ b − ∂ μ bðxÞKμ þ ; ð50Þ expect that in the case of inflation the (temporal component
2 3 96κ of the) anomaly current Kμ would be completely washed
where the … denotes higher-derivative terms, including out at the end of inflation, as a result of the exponential
higher-curvature terms irrelevant for our purposes expansion of the scale factor during the inflationary phase:
here.10 From Eq. (50), we observe that the equations of
aðtÞ ∼ expðHtÞ; ð54Þ
motion for the KR-axion field are the same as those
obtained from the action (28), i.e., they still assume the where H ≃ const denotes the (approximately) constant
form (45), but, now, the total “matter” stress tensor, for the Hubble parameter during inflation (in units in which
fields φðxÞ and bðxÞ, reads today’s scale factor a0 ¼ 1, which are used throughout).
However, as we shall demonstrate now, this is not always
T φþb
μν ¼ ∂ μ φ∂ ν φ þ ∂ μ b∂ ν b the case. Indeed, it is possible to consider scenarios
displaying cosmological birefringence during inflation.
1 1
− gμν ∂ α φ∂ α φ þ ∂ α b∂ α b − UðφÞ ; ð51Þ This means that one can distinguish the effects from chiral
2 2 gravitational components having different dispersion rela-
tions, which explains the name. In what follows, we shall
where the reader is reminded of the fact that the anomaly
explore situations in which, due to the above phenomenon,
terms do not contribute in a FLRW space-time, assumed
the right-hand side of Eq. (48) might be nonvanishing, and,
on average (however, see below, where we consider
as we shall discuss, under certain circumstances to be
8
specified below, the washing out of the anomaly triggered
In some supergravity models, the role of the inflaton might be by inflation could be avoided.
played by the real part of a complex scalar field, which represents To this end, let us consider a spatially flat FLRW space-
the dilaton Φ [see Eq. (14)], whose imaginary part is the axion; a
slow-roll dilaton, upon assuming appropriate potentials, leads time, with scale factor aðtÞ, perturbed weakly by scalar
then to inflation, and the cosmic time derivative of the axion field (ϕ̃, ψ) vector (wi ) and tensor (hij ) perturbations
might be taken to be of the same order as that of the dilaton (slow
roll for both components of the complex scalar field). This was ds2 ¼ ð1 þ 2ϕ̃Þdt2 − wi dtdxi − a2 ðtÞ
the case assumed in Ref. [52].
9
For brevity and concreteness, we assume here that the scalar × ½ðð1 þ 2ψÞδij þ hij Þdxi dxj : ð55Þ
field couples minimally to gravity. Nonminimal couplings to
gravity are certainly interesting scenarios, which however we do Only the tensor perturbations contribute to RR̃ terms, and
not consider here, as they will not be directly relevant to (in the hence we keep them in our subsequent discussion.
sense of not qualitatively affecting) the main conclusions of our
work, which are the RVM-type contributions to the vacuum Notice that the tensor perturbations constitute the non-
energy density of the Universe due to the gravitational (and diagonal part of the metric. In the study of the usual cosmic
chiral) anomalies and the novel matter-antimatter asymmetry perturbations of the matter and dark energy fields the vector
induced by the undiluted KR field at the end of inflation. part of the perturbation is set to zero and one exclusively
10
If quadratic terms in the (scalar) curvature, βR2 , β > 0, exist focuses on the Bardeen gravitational potentials ϕ̃ and ψ
in the effective action (50), then one may associate the inflaton
field φ with the scalar mode contained in those R2 terms [53], in since the nondiagonal spatial part decouples from the rest in
which case the potential UðφÞ is that of the Starobinsky model for the form of gravitational waves propagating in the FLRW
inflation [54]. background. Here, however, we rather focus on the tensor
045001-12
part and ignore the rest since it has no impact on our when the cosmic time increases the conformal time
considerations. In fact, it is only during the inflationary decreases. This is to be taken into account in the integration
stage that the primordial gravitational waves can provide a limits of each variable. Thus, the infinite future in conformal
significant contribution. After inflation they are washed out time is attained in the limit η → 0. In Eq. (59), we used the
only to reappear in the very late universe but in a much notation of Ref. [52] for the (dimensionless) quantity Θ
weaker form, as we shall see in Sec. IV. associated with the anomalous interactions in Eq. (25):
In Ref. [52], the right-hand side of the averaged rffiffiffi
Hirzebruch-Pontryagin density (48) has been evaluated 2 α0 κ _
for metrics representing gravitational-wave space-times Θ¼ Hb̄: ð61Þ
3 12
during inflation, which is a solution of Einstein’s equations
in the action (28) with the anomalous term, and we use it here At this point, we make the important remark that the nontrivial
as a prototype for yielding nonzero anomalies of relevance to result (59) induced by the (primordial) gravitational-wave
us. Assuming, for concreteness, gravitational waves propa- perturbations will imply a nonzero result on the right-hand
gating along the z spatial direction, we consider the metric side of Eq. (37), which produces a gravitational anomaly, in
the sense that the matter stress-energy tensor is no longer
ds2 ¼ dt2 − a2 ðtÞ½ð1 − hþ ðt; zÞÞdx2 þ ð1 þ hþ ðt; zÞÞdy2
conserved and, for constant G, it implies the violation of the
þ 2h× ðt; zÞdxdy þ dz2 ; ð56Þ Bianchi identity. Ultimately the reason for this situation is
that, since quantum graviton fluctuations are invoked in the
in the usual notation for the polarization of the gravitational computation, there is no guarantee that the classical Einstein
waves. For an inflationary space-time the scale factor has the equation (36) will continue to hold, and this is implied here by
exponential form (54). The CP violation, induced by axion- the nonconservation of the classical KR-axion stress tensor.
like couplings to the Hirzebruch density (48) in Eq. (25), can
Finally, we note that the nonvanishing of Eq. (59) is due to the
be seen if one uses the chiral graviton basis:
fact that inflation produces a violation of the CP symmetry
1 out of equilibrium, and this fulfils Sakharov’s necessary
hL;R ¼ pffiffiffi ðhþ ih× Þ; ð57Þ
2 conditions for baryogenesis, which will have implications for
our subsequent discussion on the generation of matter-
where the − (þ) sign pertains to L (R), and hL;R are scalar antimatter asymmetry in our model, in Sec. III B.
complex-conjugate fields. The CP-violating topological Above, we assumed slow roll for b̄,
interactions of the axion field in Eq. (25) imply inequivalent
b̄_ ≪ H=κ;
behavior of hL;R in the inflationary space-time.
ð62Þ
Taking into account that [52] Rμνρσ R̃μνρσ ≃
4ia−3 ½∂ 2z hR ∂ t ∂ z hLþa2 ∂ 2t hR ∂ t ∂ z hLþ 12 ∂ t ða2 Þ∂ t hR ∂ t ∂ z hL −
so that jΘj ≪ 1, which justifies neglecting OðΘ3 Þ terms in
ðL ↔ RÞ, which is quadratic in the graviton perturbations,
Eq. (59) [52] (the reader should recall that, during inflation,
we may make the following approximation, to leading (up
the Hubble parameter H is assumed to be approximately
to second) order in small perturbations, at which we shall
constant). This necessitates an α0 ¼ 1=M2s , where Ms is the
be working in this article:
string mass scale, such that α0 H2 ≪ 1 during inflation, for
pffiffiffiffiffiffi pffiffiffiffiffiffi
h −gRμνρσ R̃μνρσ i ≃ −ghRμνρσ R̃μνρσ i; ð58Þ which the scale factor aðtÞ appearing in Eq. (59) assumes
the de Sitter form (54). A natural choice, which we adopt in
which implies that one should use the unperturbed infla- this work, is to assume large string mass scales M s near the
tionary metric [with scale factor (54)] inside the metric reduced four-dimensional Planck mass scale, i.e.,11
pffiffiffiffiffiffi
determinant −g on both sides of Eq. (48).
The average of Eq. (48) over such a space-time then, up α0 ∼ κ 2 ¼ M−2
Pl ; ð63Þ
to second order in fluctuations hL;R , has been performed in
Ref. [52], with the result given that the inflationary Hubble scale is expected from
Z phenomenology [1] to be Hκ < 10−4 (we use here bounds
16 d3 k H2 4
hRμνρσ R̃μνρσ i ¼ 4 κ 2 k Θ þ OðΘ3 Þ; ð59Þ for single-field inflation models). Here we take for con-
a ð2πÞ3 2k3
creteness H in the range
to leading order in kη ≫ 1, where k is the standard Fourier
scale variable, and η is the conformal time defined as [52] H
∈ ½10−5 ; 10−4 Þ: ð64Þ
M Pl
dt 1
dη ¼ ⇒ η ¼ expð−HtÞ ð60Þ
aðtÞ H
In general [32], the string scale α0 is an independent
11
and in the last relation we took into account Eq. (54), which is parameter from the four-dimensional Planck scale κ2 . We shall
valid during inflation. We should note that dη and dt actually discuss the phenomenology of this more general case briefly later
have opposite signs for the inflationary solution, and hence on in the article and in the Appendix.
045001-13
From Eq. (47), then, the slow-roll conditions on b̄ðtÞ, Eq. (62), where we used Eq. (60) to write ln ðHηÞ ¼ −Ht and
pffiffiffiffiffiffiffiffiffiffiffi
should also characterize hK0 i, as a consistency check. Eq. (54) to express 1= −gðtÞ ∼ a−3 ðtÞ ¼ exp ½−3HtðηÞ
While staying in the FLRW frame, it is convenient to pass so as to integrate this expression as part of the exponential.
into conformal time η [Eq. (60)] to study the solutions of Finally, as already mentioned, we have set the beginning of
Eq. (48). We also use an ultraviolet cutoff μ for the modes, inflation at t ¼ 0 (η ¼ H−1 ), which is assumed immedi-
such that their physical momentum k=a is cut off by [52] ately after the big bang, and its end at t → þ∞ (η → 0).
The value K0begin ðtðη ¼ H−1 Þ, which on account of
kη < μ=H: ð65Þ Eq. (47) corresponds to an initial condition for the cosmic
Indeed, let us note that the leading contributions to the time derivative of the KR axion, b̄_ ð0Þ, is a boundary
momentum k integral on the right-hand side of Eq. (59) condition to be determined phenomenologically, as we
come from modes 1 ≪ kη < μ=H [52]. On using Eqs. (47) shall discuss later on. In our normalizations (60), the initial
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
and (61), and taking into account that η runs in the opposite scale factor aðtðH ÞÞ ¼ 1, and thus −gðtðH−1 ÞÞ ¼ 1.
−1
direction as the cosmic time t, we obtain from Eq. (59), to The reader should compare Eq. (68) with Eq. (49). The
leading order in the CP-violating quantity Θ [Eq. (61)], presence of gravitational waves during the inflationary
phase may lead to a decrease in general, or even complete
μνρσ 1 H 2 4
hRμνρσ R̃ i ¼ 2 μΘ elimination, of the exponential washing out effects of
π M Pl inflation as t → þ∞. Indeed, the factor A in the exponent

2 1 H 3 μ 4 on the right-hand side of Eq. (68) reads
¼ 2 MPl × K0 ðtÞ:
3π 96 × 12 M Pl MPl
−5 H 2 μ 4
ð66Þ A ≡ 1 − 1.95 × 10
MPl M Pl
4
Using this result, then from Eqs. (48), (59) and (60) we H 2 μ
get [40] ¼1− 0.664 : ð69Þ
MPl 10M Pl
d pffiffiffiffiffiffi 0 d pffiffiffiffiffiffi Due to the slow running of H during inflation, A is
ð −gK ðtðηÞÞÞ ¼ −ðηHÞ ð −gK0 ðtðηÞÞÞ approximately constant. In inflationary scenarios where
dt dη
H ≪ M Pl [Eq. (64)], and taking into account that a natural
H 3 μ 4
¼ 5.86 × 10 −5
MPl range of the cutoff μ is μ ≲ M Pl , one would expect, in
MPl MPl general, A ≃ 1, in which case the anomaly would be
pffiffiffiffiffiffi
× ð −gK0 ðtðηÞÞÞ: ð67Þ washed out at the end of inflation t → þ∞. However,
one observes that
The slow-roll nature [Eq. (62)] of K0 ðtÞ, follows immedi-
ately from Eq. (67), already from the beginning of inflation
½cf:Eq:ð69Þ H M Pl 2
A¼0 ⇒ ¼ 15.06 : ð70Þ
t ¼ 0 [or equivalently η ¼ H−1, cf. Eq. (60)], as a conse- MPl μ
quence of the fact that during inflation H ≪ M Pl
[cf. Eq. (64)]. This is a self-consistency check of our If one insists on phenomenologically acceptable ranges of
approach in adopting the solution (47). The end of inflation H ≪ M Pl , e.g., Eq. (64), then we observe from Eqs. (69)–
occurs for t ≫ M −1
Pl , and for all practical purposes we set it
(70) that trans-Planckian modes should be necessarily
here formally at t → ∞ [i.e., for conformal time (60) involved to ensure that the factor A ¼ 0, since the cutoff
η → 0]. Thus, in conformal time units the duration of in that case should exceed the Planck scale
the inflationary period is Δη ∼ H−1 .
On assuming that H remains approximately constant μ ∼ 103 MPl : ð71Þ
R η 0 the inflation period, Eq. (67) can be integrated over

during This provides, through Eq. (47), a self-consistent and
0 dη . With the above in mind, we can estimate from necessary condition for b_ to be approximately constant
Eq. (67) that during inflation, which implies a spontaneous violation of
1 the Lorentz symmetry by the KR background. The scale μ
K0 ðtðηÞÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi K0begin ðtðη ¼ H−1 ÞÞ of this violation (71), being trans-Planckian, does not affect
−gðtðηÞÞ the effective potential of the low-energy effective field

H 2 μ 4 theory at inflation, the latter defined for modes below the
× exp −5.86 × 10−5 lnðHηÞ Planck scale.
M Pl MPl
Having said that, we remark that the appearance of trans-
∼ K0begin ðtðη ¼ H−1 ÞÞexp½−3HtðηÞð1 − 1.95 × 10−5 Planckian modes, might indicate to many a potential

H 2 μ 4 breakdown of an effective field theory, or the weak gravity
× conjecture, i.e., that the effective quantum field theory we
MPl M Pl
are dealing with cannot be consistently coupled to the full
≡ K0begin ðtðη ¼ H−1 ÞÞexp½−3HtðηÞA; ð68Þ quantum gravity if Eq. (71) is valid. We, however, adopt a
045001-14
rffiffiffi
different interpretation, in that Eq. (71) offers a sign that 2 α0 0
these gravitational waves are indeed of quantum-gravity b̄_ ¼ K ≃ const: ð72Þ
3 96κ
origin and are generated deep in the trans-Planckian region
but appear to us as classical gravitational waves below the We can now use the above result to provide a phenom-
Planck scale, which is the only region we can deal with at enologically consistent estimate of K0begin ðt ¼ 0Þ. In prin-
the semiclassical level.12 In this respect, we also mention ciple, without details of the model for inflation it is not
that trans-Planckian values of the inflaton field are also possible to do this. The KR field is an independent field
considered in inflationary scenarios, but still a classical from the inflaton φ, and thus in principle, although both are
general-relativity treatment applies in such cases [55].13
The relations (71) or (A5), provide, through Eq. (47), the slow running, the only constraint is that b̄_ has to be much
self-consistent and necessary conditions for b_ to be smaller than jUðφÞj, in order not to upset the inflationary
approximately constant during inflation, which thus condition (53). A reasonable scenario, which allows a self-
remains undiluted at the end of the inflationary period of consistent phenomenology, is to assume that these two rates
the string Universe: are of the same order of magnitude. Such a case character-
izes, for instance, the scenario of Ref. [52], inspired by
string-inspired conformal supergravity models, where the
12
Nonetheless, trans-Planckian values for the cutoff μ can be axion is just the imaginary part of a complex scalar field,
avoided in the more general case, where α0 ð¼ M−2 2 −2
s Þ ≠ κ ð¼ M Pl Þ, whose real part is the dilaton. In our case, the KR axion
upon appropriately restricting the range of values of α0 , as
explained in the Appendix. originates from the same gravitational multiplet of strings
13 as the graviton and dilaton, and thus the above assumption
Nonetheless, we should remark at this point that, independ-
ently of our considerations here, it was pointed out in Ref. [56] is also reasonable. Taking into account the phenomeno-
that the predictions of Ref. [52] for leptogenesis due to primordial logical value for the slow-roll parameter for (single-field)
chiral fermions depend heavily on the ultraviolet completion of
the theory, in our case the full string theory, given that mainly
inflation ϵ, as inferred from cosmological CMB observa-
modes in the deep quantum-gravity/string-theory regime con- tions [1], we then write
tribute to the lepton asymmetry; moreover, as argued in
Refs. [56,57], by performing proper ultraviolet regularization, 1 1 1 1
b̄_ ∼ 10−2 ;
2
including higher-than-quadratic-order derivative terms, one may ϵ¼ φ_ 2 ∼ ð73Þ
effectively obtain a much smaller lepton number than the one 2 ðHMPl Þ2 2 ðHM Pl Þ2
claimed in Ref. [52], since the cutoff μ is effectively replaced by
the Hubble constant during the de Sitter phase. In contrast, in our which implies14
approach, there are no primordial fermions, and leptogenesis
during the radiation era occurs in a completely different way pffiffiffiffiffi
[37,38] to be discussed in Sec. III B, due to the presence of a b̄_ ∼ 2ϵMPl H ∼ 0.14M Pl H; ð74Þ
constant Lorentz-violating axial background of the KR field. The
latter is induced by the gravitational anomaly (59), and, as we
shall show below, remains undiluted at the end of inflation,
which can be integrated to give
provided trans-Planckian modes (71) are included. Thus, pffiffiffiffiffi
although the induced CP violation, required for a nonzero b̄ðtÞ ∼ b̄ð0Þ þ 2ϵMPl Ht; ð75Þ
(average) value of the gravitational anomaly, and thus lepto-
genesis, is generated by gravitational waves, and one needs the
full string/quantum gravity theory to determine the initial value of where b̄ð0Þ is an initial value of the KR axion field, at the
the KR axion at the big bang (t ¼ 0), nevertheless, the low- beginning of inflation, immediately after the big bang. We
energy effective field theory approach suffices for a description of shall come back to the phenomenology of this initial value
the generation of a lepton asymmetry during the radiation epoch.
As we shall discuss in Sec. III B below, the latter is proportional later on, in Sec. II C.
to the KR axion background itself, whose value at the exit from This allows, through Eqs. (47) and (63), to express the
the inflationary era is treated as a phenomenological parameter in (approximately constant, during inflation) anomaly K0 ∼
our scenario, since an exact prediction would depend on the Kbegin ðt ¼ 0Þ as [40]
details of the underlying microscopic (nonperturbative) string
theory model, which, at present, are not known. Our consid-
erations therefore are different from those of Refs. [56,57], in K0 ∼ Kbegin ðt ¼ 0Þ ∼ 16.6HM2Pl : ð76Þ
that, in our model, the lepton asymmetry can be computed in
terms of the gravitational-anomaly-induced (Lorentz-violating) From Eqs. (52) and (73), then, we can express the
KR axion background (in fact, the reader can easily verify that
such backgrounds also constitute solutions of the axion equations contributions of the anomaly to the energy density of
of motion of the one-loop effective action of Ref. [57], but no the string-inspired Universe as
predictions on their magnitude can be made in that framework,
given that the coefficients of the various terms can only be
computed if the UV-complete theory is known). Incidentally, for In Sec. III, we shall see that such an order of magnitude for ϵ,
14
a connection of the trans-Planckian problem to Lorentz violation, or equivalently b̄_ at the end of inflation (74), also leads to
but from a rather different perspective than ours, see also phenomenologically acceptable leptogenesis in the radiation era,
Ref. [58]. according to the mechanism of Ref. [38].
045001-15

φþb H 2 UðφÞ (Einstein-type) equations with respect to the gravitational
ρ ≃ 3M4Pl 3.33 × 10−3
þ : ð77Þ
M Pl 3M 4Pl field. In general this is a complicated process, where the full
string theory (or UV-complete quantum gravity) dynamics
Inflation occurs as long as UðφÞ ≫ 10−2 ðHM Pl Þ2 . The plays a role, and at present a complete formal treatment is not
terms depending explicitly on H in Eq. (77) constitute available. Nonetheless, for our purposes here, we adopt a
running-vacuum-like corrections [6] to the classical infla- phenomenological approach, in which we postulate the
tionary (almost constant) potential U. In case, for instance, existence of this condensate in a low-energy effective action
the inflationary potential is that of Starobinsky, with framework, basing this assumption on our previous results
parameter β, which arises naturally in string-inspired on the induced anomaly by means of primordial gravita-
models that contain higher-curvature corrections in their tional-wave perturbations of the de Sitter background
effective low-energy actions, the dynamical vacuum model during inflation (59). We assume that such primordial
energy density assumes the form [26] gravitational-wave perturbations set the dominant scale
for the condensate.

H 2 Once such a condensate is formed we may expand the
ρRVM ðHÞ ¼ 3M 4Pl c0 þ ν þ βH4 ; β > 0: ð78Þ gCS term (30) in the effective action (25) over quantum
MPl
fluctuations about it, by writing formally
As we can see, this expression is of the generic running rffiffiffi Z
vacuum form (5) that we have studied in the previous 2 α0 pffiffiffiffiffiffi
S b-grav
¼ d4 x −gðhb̄ðxÞRμνρσ R̃μνρσ i
section. In our case, 3 96κ
ν ∼ 3.33 × 10−3 ≪ 1; ð79Þ þ ∶bðxÞRμνρσ R̃μνρσ ∶Þ; ð80Þ
and c0 ≪ ðMHPl Þ2 may be considered as part of UðφÞ so we where ∶…∶ denotes proper quantum ordering of (quantum
field) operators, which, in the path-integral language, is
can safely ignore it when we talk about quantities during
interpreted as indicating terms with the appropriate sub-
the inflationary era. The neglected term resurfaces of
traction of the UV divergences, via regularization by means
course in the late universe and becomes the leading
of the UV cutoff μ. This quantum-ordered term can give
contribution to the DE.
rise (via its variation with respect to the gravitational field)
to a quantum-ordered Cotton tensor (34), which is traceless
C. Anomaly-induced inflation through running vacuum [cf. Eq. (35)].
In this section we wish to discuss in some detail what The reader should note the fact that, as typical with
was already announced at the beginning of the previous condensates in field theory, the quantity hb̄ðxÞRμνρσ R̃μνρσ i
subsection, namely the fact that the scalar field φ that we does not depend on the metric tensor, which thus leads to
introduced there need not be a fundamental external the addition of a DE-type term in the effective action (25),
inflaton but it can be identified with the field ϕ (different which describes the effects of the gravitational anomaly
from φ) that defines the scalar field representation of the condensate:
RVM in its full fledged form (5) or (78). This form contains rffiffiffi Z
both H2 and the higher power H4, the latter being essential 2 α0 pffiffiffiffiffiffi
to trigger inflation in the RVM. In what follows we wish, SΛ ¼ d4 x −ghb̄Rμμρσ R̃μνρσ i
3 96κ
first of all, to note that our gravitational anomaly frame- Z
pffiffiffiffiffi b̄ð0Þ pffiffiffiffiffi
4 pffiffiffiffiffiffi 7
work actually predicts the full RVM form of the vacuum ≃ d x −g 5.86 × 10 2ϵ þ 2ϵN H4
energy density, in which the higher power H 4 is generated M Pl
Z
by the gCS anomaly term, that is, the last term on the right- pffiffiffiffiffiffi Λ
≡ − d4 x −g 2 : ð81Þ
hand side of the string effective action (25). κ
This comes about upon averaging such an effective action
over the inflationary space-time, i.e., when we consider the Above, the symbol ≃ indicates an order-of-magnitude
vacuum expectation value of hbðxÞRμνρσ ðxÞR̃μνρσ ðxÞi in the estimate, and we used Eqs. (67), (71), (72) and (75),
inflationary background. This is viewed as a condensate of and took into account that Ht is bounded from above by
graviton fluctuations, which is formed in the context of a ðHtÞmax , a maximum order of magnitude, evaluated at the
UV-complete theory of quantum gravity, such as string end of the inflationary period, for which ðHtÞmax ¼ Htend ∼
theory in the present example. From a formal point of view, N ¼ 60–70, where N is the number of e-foldings. We also
such condensates appear dynamically by first averaging the set ϵ ∼ 10−2 , as required by inflationary phenomenology
(quantum gravity) partition function corresponding to the [cf. Eq. (73)]. In a sense, the term (81) is equivalent to a
low-energy effective action (25) over gravitational pertur- quantum-gravity-induced “trace” of the Cotton tensor,
bations about a de Sitter background, and then looking for which, as we have seen above, is classically traceless
local minima of this action, characterized by semiclassical [Eq. (35)]. Such a Λ-type term cannot arise in a classical
045001-16
rffiffiffi
general-relativistic treatment, and, hence, it was not con- 2 α0 κ 1 4 4
sidered in the analysis of Ref. [41]. In fact, such a vacuum C00 ≃ −ϵ μ H < 0; ð84Þ
3 192 π 2
expectation value acts as a new effective (induced) con-
tribution to the vacuum energy density (52). where we used Eq. (74) but keep the slow-roll parameter ϵ
We next notice that, if we consider b̄ð0Þ < 0 and trans- generic for the moment. From Eq. (36), this contributes to
Planckian values for jb̄ð0Þj ≫ M Pl (in analogy with what the energy density of the vacuum a negative term,16 in a
happens with the inflaton field in conventional large-field similar spirit to the Gauss-Bonnet-dilaton coupling [51],
inflationary scenarios), then the order of magnitude of the also appearing in string-effective actions, which, like the
quantity Λ > 0 in Eq. (81) does not change during the gravity-anomaly term (30), also involves terms quadratic in
entire inflationary period, for which H ≃ const, and thus it the Riemann curvature tensor:
can be approximated by a constant. In fact, for this purpose, rffiffiffi 4
it suffices to assume 2 α0 00 2 1 μ
ρgCS
¼ C ≃− 2 ϵ H4
3 12κ 3 π × 192 × 12 Ms
pffiffiffiffiffi 4
jb̄ð0Þj ≳ 2ϵN MPl ∼ 10MPl : ð82Þ −5 μ
≃ −2.932 × 10 ϵ H4 < 0: ð85Þ
Ms
Hence, the term (81) behaves as a positive-cosmological-
constant (de Sitter) type term, which is responsible for Using Eq. (70), we then obtain in order of magnitude17
inducing inflation. Quantum fluctuations of the condensate
are then responsible for deviations from scale invariance, ρgCS ≃ −1.484ϵM2Pl H 2 : ð86Þ
providing a novel mechanism for cosmological perturba-
tions to be explored further and compared with data in a From Eq. (39), and the first equality of Eq. (83), we also
future work. obtain
We would now like to demonstrate the role of the
d b 4 gCS
anomaly-condensate-induced dark energy density (81) in ðρ þ ρ Þ þ 3H ð1 þ wb Þρ þ ρ
gCS b ≃0
ensuring that the temporal (00) component of the conserved dt 3
modified stress-energy tensor T̃ μν
bþgCSþΛ [Eq. (39)], which 2
⇒ ρb ≃ − ρgCS ; ð87Þ
would correspond to the total energy of the system, is 3
positive, thus implying stability. To this end, we consider
Eq. (39), and assume a nonzero vacuum expectation value where the last result holds if dtd ðρb þ ρgCS Þ ≃ 0 and we took
(59) of the anomaly term, due to gravitational waves, and into account that the equation of state of the pure b fluid is
an isotropic and homogeneous temporal component of the wb ¼ 1, as follows from Eq. (31). Thus, we see from
Cotton tensor C00 ðtÞ. Anticipating the latter to be propor- Eq. (87) that the negative value of ρgCS is essential for the
tional to Θ2 ≪ 1 [cf. Eq. (59)], one obtains from Eq. (37)15 consistency of the approach, since it is only then that the
energy conservation of the total stress-energy tensor (39)
d 00 1 leads to consistent results, given the positivity of ρb . From
Cμ0 ;μ ¼ C þ 4HC00 ≃ − b̄_ hRαβγδ R̃αβγδ i Eqs. (86) and (87) we then obtain
dt 8
rffiffiffi
0
1 2 α κ 1 H 2 4 _2 ρb ≃ 0.9895ϵM 2Pl H2 : ð88Þ
≃− H 2 μ b̄ ; ð83Þ
8 3 12 π MPl
The KR axion stress tensor T μν
b [Eq. (31)] in Eq. (36), on
in a mean-field approximation, to lowest order in a the other hand, will contribute H2 terms to the vacuum
perturbative Θ expansion, whereby on the left-hand side energy density but of the same order of magnitude as the
of the equation we considered a (spatially flat) FLRW ∼H4 terms of the gravitational anomaly, due to Eq. (87):
background space-time. In arriving at Eq. (83) we used
Eq. (35). We also remind the reader that the notation b̄ 16
For the benefit of the reader, we note that the negativity of
denotes the KR background, satisfying Eq. (72). Assuming 00
C is robust against a change of signature of the coefficient of the
a (approximately) constant-in-time C00 and homogeneity gCS term in Eq. (25), given that the latter will be compensated by
and isotropy (i.e.. setting C0i ¼ 0) we find from Eq. (83) the a corresponding change of signature of the averaged anomaly
consistent solution (59), which is proportional to that coefficient.
17
An important remark we would like to make is that the
condition (70) is assumed to be valid as an order-of-magnitude
15
For the remainder of this subsection we treat M s ¼ ðα0 Þ−1=2 estimate, and does not imply that the cutoff μ varies with H as
and M Pl ¼ κ −1 as independent (see the Appendix) in order to H−1=2 . The quantity μ is independent of H and constant in time.
demonstrate the generic nature of our results regarding the role of This implies that the gCS term varies as H 4 , in contrast to the ρb
the condensate; we revert back to the special case α0 ∼ κ, adopted term that varies as H2 . However, for our solution under which
in our work so far, from the next section onwards. Eq. (70) is valid, both terms are of the same order of magnitude.
045001-17
1
ρb ¼ ðb̄_ Þ2 ≃ ϵM 2Pl H2 ;
potential for the bðxÞ field, as we shall discuss
ð89Þ
2 in Sec. V.
(ii) As a result of the negative contributions of the
where we used the first equality in Eq. (74). Comparing Cotton tensor to the energy density ρtotal, the
with Eq. (88) we can then see the consistency of our coefficient of the H2 terms in Eq. (91) would imply,
approach, for every value of the slow-roll parameter ϵ < 1 on account of Eq. (5), a ν < 0 in the early Universe,
and every value of H. We can then adopt the range of values where gravitational-anomaly contributions domi-
for these parameters dictated by the data [1], Eqs. (73) and nate. However there is no contradiction with the
(64), respectively. The 1% discrepancy between Eqs. (89) spirit of the RVM. Indeed, in our case, the Cotton
and (88) is to be expected, according to our discussion in tensor is not a vacuum contribution, as it is asso-
the Appendix [cf. Eq. (A4)], which implies that the result ciated with gravitational-wave excitations of the
FLRW metric background space-time. For the back-
(86) for ρgCS should be multiplied by an uncertainty factor ground space-time, the Cotton tensor vanishes, as
ξ ξ
ð1 − 3N Þ in the range 0.9889 ≲ ð1 − 3N Þ ≲ 0.9905. This is we have already mentioned [41]. On the other hand,
perfectly justified when also taking into account theoretical the KR axion is associated with the spin-1 anti-
uncertainties in our estimate (59) of the gravitational- symmetric tensor field of the massless gravitational
anomaly condensate. multiplet of strings [33], which in the case of the
We now remark that, as follows from Eqs. (85) and (87), (phenomenologically relevant) superstring consti-
the sum of the respective energy densities turns out to be tutes the ground state, due to the absence of tachyon
negative modes from the spectrum. In this sense, the RVM
should be associated with the contributions of the
1 b-axion field stress tensor T μν b [Eq. (31)] alone,
ρb þ ρgCS ¼ ρgCS ≃ −0.496ϵM2Pl H2 < 0; ð90Þ ignoring the Chern-Simons terms, which, on ac-
3
count of Eqs. (73)–(74) leads [cf. Eq. (5)] to a
positive ν coefficient (79), as mentioned previously.
indicating that, if there were no other contributions to the In the radiation- and matter-dominated eras, where
energy density of the KR axion-gravity system, the gravi- the gravitational anomalies cancel [40], as we shall
tational anomaly would induce an instability in the de Sitter discuss in Sec. III A, this is also the case.
vacuum. (iii) On the other hand, we find that the coefficient α is
However, as already mentioned, the term (81) in the positive already during the inflationary era, and of
energy density, induced by the anomaly condensate, leads order
to an additional Λ–de Sitter–type contribution to the pffiffiffi
modified stress-energy tensor (39), with an equation of 2 jb̄ð0Þj HI 2 jb̄ð0Þj
state ρΛ ¼ −pΛ , which does not modify its conservation α¼ ×5.86× ×106 ∼2.8×10−2 ;
3 M Pl M Pl MPl
(39), but corresponds to a (positive) contribution to the total
pffiffiffiffiffi ð92Þ
energy density ρΛ ≃ 5.86 × 107 2ϵ jb̄ð0Þj 4 −2
MPl H . For ϵ ∼ 10 ,
N ¼ Oð60–70Þ and jb̄ð0Þj ≳ 10MPl , it dominates the total assuming a (typical) Hubble parameter HI during
energy density, inflation of order (64). Notice that the value of α does
not depend on the specific magnitude of the string
scale, but only on the ratio μ=Ms [see the discussion
Λ H 2
ρtotal ¼ρ þρ þρ ≃
b gCS
3M 4Pl
−1.7 × 10 −3
in the Appendix, and Eq. (A5)]. From Eqs. (91) and
M Pl
pffiffiffi 4 (64), then, one easily sees that we may identify the
2 jb̄ð0Þj H total energy density with a GUT-like potential
þ × 5.86 × × 106 > 0; ð91Þ
3 M Pl M Pl V ∼ M4X corresponding to an energy scale MX :
jb̄ð0Þj
which is thus positive and drives the de Sitter (inflationary) ρtotal ≃ ρΛ ∼ M4X ≃ 8.3 × × 10−10 M4Pl
space-time. MPl

Before closing the current subsection, we would like to jb̄ð0Þj 1=4
compare the expression (91) with the form of the RVM ⇒ M X ≃ 1.3 × 1016 GeV; ð93Þ
MPl
energy density (5). For the conventional RVM, the expect-
ation is that ν, α are positive [3–7]. On comparing Eq. (91) which, for jb̄ð0Þj ≳ 10MPl [cf. Eq. (82)] is in agree-
with Eq. (5), by identifying ρtotal and ρΛRVM ðHÞ, we make ment with generic RVM predictions based on GUT
the following observations for our model: models [7].
(i) In our string-inspired model for the early Universe The next point is also of crucial interest for us. The
we have c0 ¼ 0. Such a term may appear in the late quantum fluctuations of the gravitational fields that pro-
eras of the Universe, e.g., through the generation of a duce the anomaly condensate Λ term (81) could be
045001-18
described by an effective action of a composite scalar mode, quantum scalar mode encoded in the quantum fluctuations
“ϕ,” consisting of a coherent superposition of quantum of the condensate (80), which, as we mentioned above,
b-axion and graviton modes. The (gravitational) inter- needs to be computed within the proper string theory
actions among those fields, will result in self-interactions framework. That scalar condensate mode would be the true
of the condensate field, and thus an effective potential, that “vacuumon” field, which should be used in the inflationary
can in principle be computed. In practice, as the full string phenomenology of cosmological perturbations in our
theory is in operation here, such a task is currently not scenario. Hence, the true effective potential of this
feasible. In simpler situations, for instance dynamically composite “vacuumon,” properly including all the quantum
broken supergravity scenarios, a low-energy effective corrections, might be very different from the “classical”
potential of condensates of gravitino fields has been potential (97) used in Ref. [60] to describe the classical
computed in Ref. [59], and the situation resembled the RVM evolution. Nonetheless, our arguments above indicate
Starobinsky model of inflation [54], under the conditions that, in the present string-inspired RVM scenario, where
discussed in detail in that work. gravitational-anomaly condensates coupled to KR axions
Interestingly enough, the ∼H 4 behavior can be equiv- from the massless bosonic gravitational string multiplet,
alently mapped to a scalar field behavior. Such a scalar field dynamically induce de Sitter space-times, there could be
picture will be called the “vacuumon picture” of the RVM such a fully fledged vacuumon quantum field, that also
since the field ϕ is called the vacuumon [60]. To implement represents the fluctuations of the RVM and thus could be
the mapping of the RVM to the vacuumon picture one has used for the inflationary phenomenology of the model.
the following correspondence between the total density and If we use the (correct) vacuumon representation, then, its
pressure [8,9,60]: aforementioned effective potential would contain the same
information as if one used the RVM density (5). Borrowing
ρtot ≡ρϕ ¼ ϕ_ 2 =2þVðϕÞ; ptot ≡pϕ ¼ ϕ_ 2 =2−VðϕÞ; ð94Þ the correspondence formula (95) between the two pictures
we find that the slow-roll parameter for the vacuumon is
with
_
H 1 1
2 _ ϵ¼− ¼ ϕ_ 2 ≃ 10−2 ; ð98Þ
ϕ_ 2 ¼ − 2 H; ð95Þ H 2 ðHM Pl Þ2
κ
and as we can see it takes exactly the same form as for the
and where inflaton case in Eq. (73). The upshot is that the averaged
gCS anomaly term over the de Sitter space-time leads to a
3H2 _
H 3H2 a dH 2 ∼H4 contribution to the effective vacuum energy density of
V ¼ 2 1þ ¼ 2 1þ ð96Þ
κ 3H2 κ 6H2 da the RVM and there is no need to introduce any ad hoc
inflaton to trigger inflation by hand, given that inflation can
is the effective potential of the vacuumon scalar field ϕ. be entirely driven by this term [8–13].
Once we realize that the higher-order term ∼H4 of the RVM Thus, we can use the exact same fundamental fields as
density (5) can indeed be generated thanks to the gravi- the ones we started with in the effective action of bosonic
tational-anomaly term, we can just use the vacuumon string theory in Sec. II A. The RVM density (5) appears to
picture. In particular, using Eq. (96), one can compute be an effective description of the same physical context
the effective potential associated to the RVM density, when it is averaged over the inflationary space-time. Such a
whose explicit form was given in the aforementioned description can alternatively be formulated within the
references, with the result vacuumon picture and in this case it is a scalar field (the
vacuumon) which mimics the ∼H4 behavior (and thus the
H2I 2 þ cosh2 ðκϕÞ inflaton behavior) through an appropriate effective poten-
UðϕÞ ¼ : ð97Þ
ακ 2 cosh4 ðκϕÞ tial. The vacuumon, therefore, is not an external scalar field
but just an internal d.o.f. associated with the gCS anomaly,
In this scenario, if the potential (97) were the true potential leading to the scalar field representation of the higher-order
that describes the dynamics of the quantum fluctuations of ∼H4 term in the original averaged effective action over the
the scalar anomaly condensate in our case, this would be de Sitter background. This fact allows us to entirely
the potential assumed in Eq. (77). reproduce the same considerations as in the previous
However, there are some subtle issues in the approach of section but without invoking any new scalar field, which
Refs. [8,9,60] that prevent one from extending it straight- would be extraneous to our original massless bosonic
forwardly to the case examined in this article. The scalar gravitational multiplet of string theory (as this would
field ϕ in Eqs. (94) and (97) is a classical field, which is require an appropriate dilaton potential, in case the dilaton
used to describe the temporal evolution of the classical is identified with the inflaton, which however cannot be
RVM vacuum. It is by no means equivalent to the true generated at tree level in string loop perturbation theory, but
045001-19
rffiffiffi rffiffiffi rffiffiffi 2

requires higher string loops, which we do not have control pffiffiffiffiffiffi 3 5μ 2κ μ 3 e pffiffiffiffiffiffi μν
of). The RVM formulation is therefore fully self-consistent ∂ μ −g κJ − K ¼ κ −gF F̃μν
8 3 96 8 8π 2
for the description of the cosmic evolution. rffiffiffi rffiffiffi
3 e2 0ijk 3 e2 pffiffiffiffiffiffi
¼ − κ 2 ϵ F0i Fjk ¼ − κ 2 −gEi Bj gij ; ð99Þ
8 4π 8 2π
III. POST-INFLATIONARY ERA AND
ANOMALOUS MATTER OVER
where we used Eqs. (20)–(21); Ei and Bi denote the
ANTIMATTER DOMINANCE
cosmic electric and magnetic fields in curved space,
A. Chiral fermionic matter and cancellation respectively [from the third equality in Eq. (99), the
of gravitational anomalies reader can readily see the topological nature (i.e.,
At the end of inflation, the proper decay of the running independence of the metric) of the chiral anomaly,
vacuum to matter and radiation components will reheat the which thus, unlike the gravitational anomaly, does not
universe and lead to the appearance of fermions among have any contributions to the stress-energy
P tensor of the
other matter. If such fermions have anomalous axial KR axion field19]; and J 5μ ¼ j ψ̄ j γ μ γ 5 ψ j is the axial
currents, then matter-antimatter asymmetry in the observ- current, with the summation being over appropriate
able Universe could be due to such an anomaly in the post- fermion species ψ j of the matter sector, e.g., charged
inflationary era through the mechanism advocated in chiral quarks or leptons in the SM sector.
Refs. [36–38], as we now proceed to explain.18 The reader is reminded that the appearance of the square
To this end, we first assume that the space-time after of the QED coupling e (electron charge) on the right-hand
inflation has the ordinary FLRW form (in the radiation era), side of Eq. (99), is a result of the fact that the chiral
since any primordial gravitational-wave perturbations anomaly (like the gravitational anomalies) is a one-loop
would have been washed out during inflation. This would exact effect [62], with the chiral fermions circulating in the
imply that the gravitational anomaly (48) would vanish at loop. For concreteness and brevity, in Eq. (99) we assumed
large scales for such space-time backgrounds. However, the circulation of a single chiral fermion of charge equal to
locally gravitational-wave perturbations are present, and the electron charge e. In realistic applications, one should
could jeopardize the local diffeomorphism invariance of the replace e2 on the right-hand side of Eq. (99) by an
radiation (and matter) quantum theory, according to our “effective” squared charge:
previous discussion. We now postulate that the generation
of chiral matter at the end of inflation leads to a cancella- e2 ⇒ e2eff ¼ e2 N ; ð100Þ
tion of the gravitational anomalies, even locally. Otherwise
diffeomorphism invariance would be violated locally in the
presence of matter. However, and this will turn out to be where N is a model-dependent numerical constant, which
crucial for linking KR axions to DM in our scenario, as we depends on the number and kind of fermions circulating in
shall discuss later, we assume that U(1) chiral anomalies the loop, and is proportional to the square of their electric
[62] remain uncompensated. These do not contribute to the charges normalized to the electron charge e. For instance,
stress tensor of matter, unlike the gravitational ones, and for QCD chiral anomalies, of N f species of light quarks,
hence there is no fundamental reason for the matter theory with electric charges qI , I ¼ 1; …N f , each of which comes
to be chiral-anomaly free: only the gauge symmetry must in N c colors (for ordinary QCD, N c ¼ 3), one has [63]
PN f qI 2
be anomaly free so as to preserve the Ward identities. Thus, N ¼ NNfc I¼1 ð e Þ . The generalization (100) will be under-
we postulate the following relation during the radiation
stood in what follows.
(and matter) era [40]:
We stress once more that, in our approach, the U(1)
photon and fermion fields are produced by the decay of the
18
running vacuum at the end of the inflationary era [8].
We remind the reader that in our approach we do not discuss During the exit phase from inflation, there is also the KR
the role of (primordial) fermionic excitations during inflation,
since we assume that only bosonic gravitational d.o.f. describe
the string-inspired Universe. Thus the considerations of Ref. [52] 19
In this work, for simplicity, we consider only chiral U(1)
for generating sufficient leptogenesis only through the gravita- anomalies. In general, one may face situations in which there are
tional anomaly induced by gravitational waves do not apply here, also QCD triangle anomalies, which would amount to adding to
given that the relevant fermionic chiral matter in our model is the right-hand sides of the first
qffiffi and subsequent equalities of
generated only at the end of inflation, not during inflation. For 0 α pffiffiffiffiffiffi
completeness, we mention though that there are works in the Eq. (99) a term of the form þ 38 ακ ð8π
s
−gGaμν G̃aμν Þ, where Gaμν
literature [47,61] which discuss the possibility that primordial denotes the gluon field strength, where a ¼ 1; …8 is an adjoint
fermionic torsion contributions in torsional versions of general SU(3) color index, and αs is the strong interaction fine-structure
relativity (in which the spin connection and vielbein are treated as constant. This term, like the chiral U(1) anomaly one, is also
independent fields), result, through appropriate fermion conden- topological and does not yield any contributions to the stress-
sates, in inflation. We shall not discuss such scenarios here. energy tensor of the KR field.
045001-20
Z rffiffiffi
axion, which is undiluted, [Eqs. (47) and (74)]. As we shall
4 p ffiffiffiffiffiffi 1 1 μ 2κ μ
S ¼ d x −g − 2 R þ ∂ μ b∂ b −
eff
∂ bðxÞK
discuss below, this field plays an important role in both the 2κ 2 3 96 μ
cancellation of the gravitational anomaly and inducing Z rffiffiffi
leptogenesis during the radiation era [36–38]. 4 pffiffiffiffiffiffi κ 3
þ SDirac þ d x −g F μ þ
Free ∂ b J 5μ
Let us see these effects in a detailed manner by discussing 2 2 μ
the low-energy (string-inspired) effective action during the Z
3κ 2 pffiffiffiffiffiffi
radiation era. First we remark that, upon inclusion of − d4 x −gJ5μ J5μ þ þ ; ð102Þ
16
fermionic matter at the end of inflation, the contorsion
interpretation of the antisymmetric tensor field strength [33– where the … in Eq. (102) indicates gauge field kinetic
36], Hρμν , implies a minimal coupling of this field to the axial terms, as well as terms of higher order in derivatives, which
fermion current, given that the corresponding Dirac are of no direct relevance to us here. The reader should
Lagrangian for fermions in torsional gravitational back- notice the four-fermion axial-current-current term in
grounds [47,61] contains the generalized spin connection Eq. (102), which is characteristic of Einstein-Cartan the-
ω̄abμ ¼ωabμ þK abμ , K abc ¼12ðHcab −Habc −Hbca Þ¼−12Habc : ories with torsion [47,61], the latter being provided here
[33,34] by the (totally antisymmetric) quantity ϵμνρσ ∂ σ b
Z which is dual to the Kalb-Ramond antisymmetric tensor
pffiffiffiffiffiffi { field strength Hμνρ , as discussed in Sec. II [cf. Eq. (16)].
SDirac ¼ d x −g ðψ̄ j γ μ D̄ðω̄Þμ ψ j
4
2 We also remark that local gravitational-wave perturba-

tions during the radiation and matter (dust) eras lead in
− ðD̄ðω̄Þμ ψ̄ j Þγ μ ψ j Þ − mðjÞ ψ̄ j ψ j ; general to a nontrivial background F μ in Eq. (102);
Z however, such perturbations are much more suppressed
4 pffiffiffiffiffiffi { a↔
¼ d x −gψ̄ j Γ ∂ a − m ψ j during the radiation (and matter) eras as compared with
2 their primordial counterparts. In the subsequent discussion
Z
p ffiffiffiffiffiffi in this section, we consider a pure FLRW background as a
− d4 x −gðF a þ Ba Þψ̄ j γ 5 Γa ψ j
sufficient approximation of the Universe at large scales in
Z late eras. For such a pure FLRW metric gμν background
pffiffiffiffiffiffi
≡ SFree
Dirac þ d4 x −gðBa þ F a ÞJ 5a ; ð101Þ (and in general spherically symmetric space-times with
diagonal metrics [64]) one has that F a ¼ 0.
The KR axion bðxÞ background field equation of motion
with latin indices a; b; c; … denoting tangent-space indices, then, obtained from Eq. (102), reads
raised and lowered by the Minkowski metric ηab of the rffiffiffi rffiffiffi

tangent space (at a point with coordinates xμ ) of a space-time pffiffiffiffiffiffi α 2 κ α 3 5α
∂ α −g ∂ b̄ − K þ κJ ¼0⇒
with metric gμν ðxÞ ¼ eaμ ðxÞηab ebν ðxÞ, where eaμ ðxÞ are the 3 96 8
vielbeins and eμa ðxÞ their inverse. Γa is a tangent-space Dirac rffiffiffi
pffiffiffiffiffiffi α 3 e2 5
matrix, such that γ μ ðxÞ ¼ eμa ðxÞΓa , and we used the standard ∂ α ½ −g∂ b̄ ¼ κ a ðtÞEi Bj δij ; ð103Þ
↔ 8 2π 2
notation for χ̄ ∂ a ψ ¼ χ̄∂ a ψ − ∂ a χψ. The covariant deriva-
tive is defined as D̄a ¼ ∂ a − 4{ ω̄bca σ bc , σ ab ¼ 2{ ½Γa ; Γb , where, in the second line, we used Eq. (99) and the FLRW
F d ¼εabcd ebλ ∂ a eλc , Bd ¼− 14 εabc d Habc , and J5μ ¼ metric, gij ¼ a2 ðtÞδij , i, j ¼ 1, 2, 3. The alert reader should
μ 5 5a a 5
ψ̄ j γ γ ψ j , and correspondingly J ¼ ψ̄ j Γ γ ψ j. In arriving have noticed that one would have arrived at the same
equation, had one used the absence of gravitational
at Eq. (101) we used the standard properties of the flat
anomalies in a background FLRW space-time, but of
(tangent) space Γa matrices. course our result emerging from anomaly cancellation is
Adding Eq. (101) to Eq. (15), implementing the con-
more general as it is independent of any metric perturba-
straint (22) via a Lagrange multiplier pseudoscalar field
tions (such as gravitational waves) that would jeopardize
bðxÞ,20 canonically normalized as before, and integrating the diffeomorphism invariance of the radiation/matter
over the field H in the path integral, we easily arrive at an quantum field theory.
effective action [using Eq. (63)]: Nonetheless, for the purposes of our discussion in this
section, we do assume on average a FLRW space-time
20
It is crucial for the reader to notice that we keep only the during the radiation era at large scales, for which gravi-
gravitational part of the anomaly, setting the non-Abelian gauge tational-wave perturbations are suppressed. In this case, the
fields A to zero; we stress that we do not include Abelian U(1)
Chern-Simons terms in the modified Bianchi identity (19), as we chiral anomaly term on the right-hand side of Eq. (103) is
anticipate the existence of chiral U(1) anomalies only in the associated with the covariant derivative of the axial fermion
fermion sector of the model. current [62]
045001-21
1 pffiffiffiffiffiffi e2 fields, corresponding to a single mode of momentum

J5μ ;μ ¼ pffiffiffiffiffiffi ∂ μ ð −gJ5μ Þ ¼ 2 Fμν F̃μν k > 0, such that22
−g 8π
e2 2 Bðt;zÞ ¼ BðtÞð−sinðkzÞ;cosðkzÞ;0Þ;
¼− a ðtÞEi Bj δij : ð104Þ
2π 2 1_ 1_
Eðt;zÞ ¼ − Bðt;zÞ ¼ − BðtÞð− sinðkzÞ; cosðkzÞ; 0Þ:
Assuming homogeneous and isotropic situations at k k
large (cosmological) scales, we only consider cosmic- ð107Þ
time-dependent backgrounds b̄ðtÞ, hJ50 ðtÞi. We denoted
the background for the fermion axial current by h…i, as we Such configurations have been argued in Ref. [63] to play a
may also assume thermal averages [in our treatment role in providing a source for the dark energy in the
we assume the existence of chiral currents, as, e.g., is Universe. We shall take a different point of view in the
the case of the SM chiral (left-handed) leptonic current, current work, where we shall argue that such configurations
P ðfÞ ðfÞ ðfÞ can lead to a source of (stiff [42]) dark matter, through the
J5L ¼ f ðl̄L γ μ lL þ ν̄ðfÞ γ μ νðfÞ Þ, where lL ðνðfÞ Þ are the
solution (103) of the KR background.
charged leptons (active neutrinos), and f is a generation The important thing to observe [65] is that the chiral
number. In models beyond the SM, other chiral fermions anomaly corresponding to Eq. (107) has only time depend-
might play a role, as well].21 ence for a FLRW metric with a scale factor aðtÞ:
Some discussion is required at this stage concerning the
space-time dependence of the electromagnetic fields, EðxÞ pffiffiffiffiffiffiffiffiffiffiffi 1 d 2
and BðxÞ (with bold face notation referring to three vectors) −gðtÞEi ðt;zÞBj ðt;zÞgij ðtÞ ¼ −a5 ðtÞ ðB ðtÞÞ: ð108Þ
2k dt
entering Eqs. (99) and (104). It is clear that one cannot have
just time-dependent fields, since, on account of Maxwell’s In such a case, the general solution of Eq. (103) is
equations, ∇ × E ¼ −B, _ where ∇ is the spatial gradient.
rffiffiffi Z t
To have nontrivial chiral anomalies at large (cosmological) C 3 e2 1 1 d 2
scales, one may adopt the simplified (but concrete) example b̄_ ¼ 3 0 − κ 2 3 dt0 a5 ðt0 Þ ðB ðtÞÞ
a ðtÞ8 4π a ðtÞ 2k dt
considered in Ref. [65], according to which one has a rffiffiffi Z t
monochromatic configuration of magnetic and electric C0 1 3 e2 1 2
¼ 3 þ κ B ðt0 Þ dt0 aðt
_ 0Þ
a ðtÞ k 2 4π 2 a3 ðtÞ
rffiffiffi
3 e2 1
21
The expansion of quantum fermionic axial currents around C0 1
such backgrounds is performed by writing J 50 ¼ hJ 50 i þ ¼ 3 þ B2 ðt0 Þ; ð109Þ
quantum fluctuations in Eq. (102). We ignore the quantum a ðtÞ kMPl 2 4π 2 a2 ðtÞ
fluctuations for our (classical) treatment in this section. This
implies that, when we consider quadratic expressions of the axial where C0 is a constant, which we shall determine later on by
current appearing in Eq. (102) [and in the stress tensor computed using continuity requirements for the b field at the interface
from it; see below Eq. (115)] we should use between the inflation and radiation eras. To arrive at the
middle equality in Eq. (109), we took into account that the
J 50 J 50 ≃ hJ 50 i2 > 0; ð105Þ amplitude BðtÞ of the magnetic field intensity scales with
the scale factor as [66]
etc., which will be understood in what follows. However, it
should be mentioned for completeness that, when one considers Bðt0 Þ
fully quantum corrections, including fermion path integration, as BðtÞ ¼ ; ð110Þ
essential when dealing with fermions, then spatial components of a2 ðtÞ
the axial current J 5i should in general be considered in fermionic
terms, and in general one may face a situation where quantum where t0 is the age of the Universe, and, thus, Bðt0 Þ denotes
fermion condensates, ⟪J 5μ J 5μ ⟫ ≠ ⟪J 5μ ⟫⟪J 5μ ⟫, could arise, which today’s value.
could take on negative values [constant in cosmic time, for some During the radiation era, as follows from
qffiffiffiffiffiffiffiffi equations, the scale factor behaves as aðtÞ ∼
period of the (early) Universe] Einstein’s
1=2
⟪J 5μ J 5μ ⟫ < 0: ð106Þ ð2 Ωrad0 H 0 tÞ , while the Hubble parameter is given by
HðtÞ ¼ 1=ð2tÞ, with the subscript “0” indicating present-day
This can lead to inflation (in the sense of equations of state of the quantities. Hence, Eq. (109) yields
form p ≃ −ρ) in models where primordial fermions are consid-
ered [47,61]. For our purposes, where primordial fermionic
22
matter excitations are assumed to not be present in the effective The relative sign differences between Eq. (107) and the
action (50) during the inflationary era, we shall consider the case corresponding solution of Ref. [65] are due to the opposite sign
(105), where only the temporal component hJ 50 i of the axial of the term coupling the KR axion with the chiral anomaly in
current of some chiral matter is nonzero during the radiation and Eq. (102) from that of the corresponding term in the action of
matter eras. Ref. [65].
045001-22
rffiffiffi
C 1 3 e2 B2 ðt0 Þ do not contribute to the covariant stress tensor, whose
b̄_ ¼ 3 0 þ qffiffiffiffiffiffiffiffi HðtÞ : ð111Þ conservation is thus not affected, in contrast to the infla-
a ðtÞ Ωrad 2 4π 2 kMPl H0
0 tionary phase, where the gravitational anomaly is present.
Solutions to the equations (114) have been discussed in
Notice that the chiral-anomaly contributions to the KR Ref. [61]. It is important to notice that during the radiation
background field are proportional to the Hubble parameter era, the fermions, like all other matter species in the model,
HðtÞ during the radiation era. If one considered the solution are relativistic and hence, cannot be simply assumed to
with C0 ¼ 0, then such corrections would contribute purely have only temporal derivatives, i.e., spatial derivatives ∂ i ψ
H2 -running-vacuum-type corrections [Eq. (78)] to the should also be considered. This complicates the detailed
energy density [8]. However, in view of the smallness of expressions for the stress tensor. However, for our purposes
cosmic magnetic fields in the Universe, including possible here, we may simply follow the approach of Ref. [61], and
primordial ones [63,65,66], we expect such terms to be estimate that such extra contributions will simply be
suppressed compared to the a−3 ðtÞ term in the early absorbed into the energy density (and pressure) of free
Universe, when C0 ≠ 0, a case relevant for leptogenesis radiation ρrad rad
0 (p ), which dominates both the KR-axion-b
[38], as we shall discuss below. contributions and those from the self-interactions of the
At present, we note that, on using Eq. (104), for fermions induced by the axial current-current hJ50 i2 inter-
homogeneous and isotropic backgrounds, we can equiv- actions due to the H torsion.
alently write the solution (109) as On account of Eq. (113), then, the energy density for the
rffiffiffi fermions acquires the form (we ignore mass terms during
C 3 50
b̄_ ¼ 3 1 −
the radiation era, as the species are assumed to be
κhJ i; ð112Þ
a ðtÞ 8 relativistic)
rffiffiffi
where C1 ≠ C0 (in general) is another integration constant. 3κ2 5 2 3κ 2 5 2 3_ 5
T 00 ≃ ðT Dirac Þ00 −
F Free
hJ 0 i ¼ hJ0 i − κ b̄hJ i þ
For our purposes, and in the spirit of our treatment in 16 16 8 0
Sec. II, we take C1 ¼ 0, and hence 9κ 2 5 2
≃ hJ i þ ; ð116Þ
rffiffiffi 16 0
_b̄ ¼ − 3κhJ50 i: ð113Þ
8 where the … denotes pure radiation contributions from the
kinetic terms which scale with the scale factor as a−4 ðtÞ. On
Next we shall determine the (classical) energy-momen- the other hand, the energy density of the KR axion reads
tum tensor, and check the self-consistent condition to
1 3κ 2 5 2
obtain a total equation of state compatible with radiation T b00 ¼ ðb̄_ Þ2 ¼ hJ i : ð117Þ
dominance, which we used in order to arrive at the above 2 16 0
results. To this end, we first notice that, the fermion
The spatial and time-space components of T F;b
ij
equations of motion (species j), derived from Eq. (102), are
[Eq. (115)], computed from Eq. (102), are [61]
rffiffiffi
μ a ðjÞ 3 2
{ea Γ ∇μ ψ j − m ψ j þ κ ∂ b̄Γa Γ5 ψ 3κ 5 2
8 a T Fij ¼ gij p ¼ gij
F
hJ i þ ;
16 0
3κ 2
− ðψ̄ l Γa Γ5 ψ l ÞΓa Γ5 ψ j ¼ 0; ð114Þ 1
8 T bij ¼ gij pb ¼ gij ðb̄_ Þ2 ; T F;b
0i ¼ 0; ð118Þ
2
where ∇μ denotes the gravitational covariant derivative on
spinors of the species j with respect to the torsion-free where again the … denotes relativistic ∼a−4 ðtÞ contribu-
connection. One can then write the (classical) stress tensor tions from the free kinetic terms of the fermions.
for the fermions as [47,61] The total energy density ρtot and pressure ptot are then
given by
{X 3κ2
T Fμν ¼ ðψ̄ j γ ðμ ∇νÞ ψ j − ð∇ðμ ψ̄ j Þγ νÞ ψ j Þ − g J5 J5α :
2 j 16 μν α 00 ¼ ρ ¼ T 00 þ T 00 þ ρ ;
T tot tot F b rad
ij ¼ gij p ¼ gij ðT ij þ T ij þ p Þ;
T total ð119Þ
tot F b rad
ð115Þ
Above, we took into account that in the radiation/matter where the superscript “rad” denotes the conventional
phase of the Universe, in which we assume that gravita- contributions from free relativistic species in the model,
tional anomalies cancel, the topological, chirally anoma- including photons, with an equation of state prad ¼ 13 ρrad ,
lous b-axion-fermionic matter coupling terms in Eq. (102), scaling as a−4 ðtÞ.
045001-23
By comparing Eq. (116) with Eq. (118), the reader can d F d

ðρ þ ρrad Þ þ 4HðρF þ ρrad Þ ≃ ðρrad Þ þ 4Hðρrad Þ ¼ 0;
readily verify that this is also the total equation of state for dt dt
the axial current-current contributions in the fermion fluid, d b
ρ þ 6Hρb ≃ 0; ð124Þ
dt
1
pF ¼ ρF : ð120Þ
3 which provide a self-consistency check of the approach.
Continuity requires matching the background (123) with
However, as follows from Eq. (111), for C0 ≠ 0, the scaling Eq. (47) [under Eq. (63)] at the temperature just at the exit
of pF and ρF is not purely a−4 (as would be the case with of inflation, T i , which, we take to be the Gibbons-Hawking
C0 ¼ 0), but contains a superposition of terms with differ- temperature [67]
ent scalings, ∼a−6 , ∼a−5 and ∼a−4 . We would like to stress
H
that Eq. (120) is the result of the solution (113) and the fact Ti ¼ ð125Þ
that, in our string-inspired model, the KR axion is a fully 2π
fledged dynamical field.23
where H ≃ HI ∼ 10−5 MPl is the value of the Hubble
On the other hand, the KR axion component is charac-
constant during the inflationary period (64). Taking into
terized by a “stiff matter” [42] equation of state
account, then, that, during the radiation era, the temperature
(T)/cosmic time (t) relation assumes the (standard cosmol-
pb ¼ ρb : ð121Þ pffiffiffiffiffiffi
ogy) form, t ¼ 0.3 8π g1=2 −2
⋆ M Pl T , where g⋆ (assumed to
be approximately temperature independent) denotes the
but again, on account of Eq. (111), the scaling of pb and ρb
total number of relativistic d.o.f. of the model under
is not a−6 alone; each contains a superposition of terms consideration, this implies
∼a−6 , a−5 and a−4 .
On account of the conservation of the total stress tensor C00 ¼ 3.5 × 1011 M 2Pl ; ð126Þ
T tot
μν [Eq. (119)], which is respected in the presence of chiral
anomalies, as already explained, one may write where we absorbed T-independent numerical constants into
the definition of the constant C0 ⇒ C00 in Eq. (123). The
ρ_ tot þ 3Hðρtot þ ptot Þ ¼ 0 scaling of the background (123) with the temperature, then,
d F d during the radiation era, is
⇒ ðρ þ ρrad Þ þ 4HðρF þ ρrad Þ ¼ − ρb − 6Hρb ; ð122Þ 3
dt dt
_b̄ ≃ 3.5 × 1011 M2 T : ð127Þ
Pl
where we used Eqs. (120)–(121). M Pl
If one recalls that the cosmic electromagnetic fields are
expected to be suppressed [63,65,66], one may make the As we shall see in the next subsection, such backgrounds
reasonable assumption that it is the first term on the right- can produce phenomenologically correct leptogenesis.
hand side of Eq. (111) which dominates, at least during the
early stages of the radiation era, implying a scaling B. KR-axion-induced leptogenesis and
[cf. Eq. (113)] matter-antimatter asymmetry in the universe
rffiffiffi Indeed, as discussed in Refs. [36–38], the presence of the

_b ¼ − 3κhJ50 i ≃ C0 :
background (123) could lead, in principle to leptogenesis,
ð123Þ as it spontaneously breaks Lorentz, CP and CPT sym-
8 a3 ðtÞ
metry. In Ref. [38] we discussed the generation of matter-
antimatter asymmetry in the presence of backgrounds of the
On making the further physically reasonable assumption
KR field precisely of the form (127), which are considered
that it is the radiation fields that dominate over the KR
to be slowly varying during the (short) freeze-out era of
contributions in the stress tensor during the radiation era
leptogenesis, as explained in that work.
[and thus drive the scaling aðtÞ ∼ t1=2 of the Universe],
In particular, we considered lepton-number asymmetry
ρrad ≫ ρF , prad ≫ pF , one obtains a self-consistent
originating from tree-level decays of heavy sterile (right-
(approximate) vanishing of both sides of Eq. (122) sepa- handed, Majorana) neutrinos (RHN) into SM leptons. The
rately, i.e., the equations relevant part of the Lagrangian is given by
23 m
The situation should be contrasted with the corresponding L ¼ LSM þ iN̄= Bγ 5 N
∂ N − N ðN c N þ N̄N c Þ − N̄=
case of torsional space-time studied in Ref. [61], where the
X 2
equation of state characterizing the torsion-induced fermion self-
interaction contributions to the stress tensor was that of stiff − yf L̄f ϕ̃d N þ H:c: ð128Þ
matter [42] pF ¼ ρF . f
045001-24
where LSM denotes the SM Lagrangian, N is the RHN field, Channel I∶ N → l− hþ ; νh0 ;
of (Majorana) mass mN , ϕ̃ is the SU(2) adjoint of the Higgs Channel II∶ N → lþ h− ; ν̄h0 ; ð130Þ
field ϕ [where ϕ̃di ≡ εij ϕj with i, j ¼ 1, 2, SU (2) indices, is
the SU(2) dual of the Higgs field], and Lf is a lepton where l are charged leptons, ν (ν̄) are light, “active,”
(doublet) field of the SM sector, where f is a generation neutrinos (antineutrinos) in the SM sector, h0 is the neutral
index, f ¼ e, μ, τ, in the standard notation for the three SM Higgs field, and h are the charged Higgs fields, which, at
generations; yf is a Yukawa coupling, which is nonzero high temperatures, above the spontaneous electroweak
and provides a nontrivial (“Higgs portal”) interaction symmetry breaking, of interest in this scenario, do not
between the RHN and the SM sectors. In the models of decouple from the physical spectrum. As a result of the
Refs. [36–38] a single sterile neutrino species suffices to nontrivial B0 ≠ 0 background [Eqs. (129) and (127)], the
generate phenomenologically relevant lepton asymmetry, decay rates of the Majorana RHN between channels I and II
and hence from now on we restrict ourselves to the first are different, resulting in a lepton asymmetry [38],
generation (f ¼ e, setting ye ¼ y). The quantity = B ¼ γ μ Bμ
appearing in the axial current term of Eq. (128) is defined in ΔLTOT ðT ¼ T D Þ Φ
terms of the four-vector ∼q 0 ; q > 0; ð131Þ
s mN
where s is the entropy density of the Universe, T D denotes

Bμ ¼ M−1 _
Pl b̄δμ0 : ð129Þ the temperature at which this asymmetry freezes out
(“freeze-out point”), that is when the total decay width Γ
for the decays (130) equals the Hubble rate of the Universe,
It denotes the Lorentz (LV), CP (CPV) and CPT (CPTV) HðT D Þ ≃ Γ, and the quantity Φ0 is defined as [38]
violating background (127), with Bμ having only a tem- 3
poral component. For such (slowly varying in the cosmic T
B0 ðTÞ ¼ Φ0 : ð132Þ
frame) backgrounds, as in our case here, the Lagrangian mN
(128) assumes the form of a Standard Model extension
(SME) Lagrangian in a Lorentz and CPTV back- The lepton asymmetry (131) can then be communicated to
ground [68]. the baryon sector via baryon-minus-lepton-number (B − L)
At this stage we should make an important remark. As conserving sphaleron processes in the SM [39], thus
the reader should have noticed, in our model, the back- producing the observed amount of baryon asymmetry
ground (129) has a derivative form, Bμ ∝ ∂ μ b, which, by (baryogenesis) in the Universe, by requiring that the lepton
partial integration, implies a coupling of the KR axion to asymmetry (131) is of Oð8 × 10−11 Þ, as indicated by
the derivative of the axial current in the effective action (cosmological) observations [1]. The number q > 0
(128). In our model, the RHN are massive in the radiation expresses theoretical uncertainties in the analytical deriva-
epoch, where leptogenesis occurs, and hence the classical tion of the lepton number asymmetry in Ref. [38], where
axial current is not conserved, since its four-divergence the Padè approximant method was used to solve the
equals imN ðN c γ 5 N þ N̄γ 5 N c Þ, as follows from the pertinent system of coupled Boltzmann equations associ-
(Majorana) equation of motion of the free RHN fields. ated with Eq. (130). The precise value of q depends on the
Therefore, the nontrivial coupling of the KR axion to the freeze-out point. Using Eq. (127), we may write
RHN current is guaranteed, independent of any
potential anomalies, and thus is consistent with the can- 11 m3N
Φ0 ¼ 3.5 × 10 : ð133Þ
cellation of gravitational anomalies by the chiral matter in M 2Pl
the radiation- and matter-dominated eras, advocated in our
scenario.24 By demanding phenomenologically acceptable values of
In the context of the model (128), a lepton asymmetry is the lepton asymmetry (131) of order Oð8 × 10−11 Þ, one can
generated due to the CPV and CPTV tree-level decays of then infer from Eq. (133) that
the RHN N into SM leptons in the presence of the
background (129) [36–38]: 1.5 3.7
mN ≃ pffiffiffi × 10−11 MPl ≃ pffiffiffi × 107 GeV: ð134Þ
q q
24
Gravitational anomalies may play a role in a dynamical
generation of the RHN (Majorana) mass, as in the scenario of The reader should bear in mind that in the semianalytic
Ref. [43], involving mixing of the KR field with other string method of Ref. [38] only the following combination of
theory axions. Such mechanisms can be consistently embedded
in our framework, specifically in the early radiation epoch, just parameters, involving mN , enters the series expansions of
after inflation, when chiral matter is generated. However, their the solutions about a point x ¼ mN =T used to approach (via
discussion falls beyond the scope of the current work. Padé approximants) the freeze-out point xD ¼ 0.1:
045001-25
MPl that is, assuming no new physics (such as supersymmetry)

I ≡ y2 : ð135Þ
mN at the TeV scale. Indeed, the Higgs-mass-squared correc-
tions coming from one-loop contributions in the electro-
We now notice that the ratio y2 =mN appears in the weak-symmetry-broken phase, due to the Higgs portal
expression for the SM active neutrino ν masses via the interactions in Eq. (128), involving (in their generality)
(type-I) seesaw mechanism [69],25 the three active and three sterile
P neutrinos
P3 propagating in
2 1 2 2
the loop, read [70] δmH ¼ α¼e;μ;τ I¼1 ð4πÞ2 yαI mI . In
mν ∼ jyj2 v2 =mN : ð136Þ order to ensure Higgs mass stability/naturalness, then, one
must have δm2H ≲ m2H , where mH is the Higgs mass. In our
In Ref. [38], the Yukawa coupling y ∼ 10−5 and mN ∼ single sterile neutrino case, considered above, we may
105 GeV [36–38] gave phenomenologically relevant eliminate the Yukawa coupling y, using the type-I seesaw
values for mν. Such parameters correspond to formula (136), to obtain the following criterion for mass
[cf. Eq. (135)] stability:
I ∼ 103 ð137Þ mN ≲ ðm2H v2 ð4πÞ2 m−1

ν Þ
1=3 : ð139Þ
which we keep fixed in our approach, so that the Using the cosmological bound
considerations of Ref. [38] apply, and moreover one P3 [1] for the sum of the three
active neutrino masses i¼1 mνi < 0.12 eV, and trans-
obtains the same (phenomenologically consistent) active lating it (on account of the neutrino oscillation data on the
neutrino masses via the seesaw mechanism as in active neutrino mass differences, assuming normal or
Ref. [38] [cf. Eq. (136)]. inverted hierarchies [71]) into an upper bound for the
Additionally, the assumption that T D ≃ mN was made in single active neutrino we consider here, mν ≲ 0.04 eV, we
Ref. [38], which we also maintain here. In such a case [38] may replace the mν in Eq. (139) by this upper bound, to
q ¼ Oð10Þ, and from Eq. (134) one obtains obtain a sufficient condition for the satisfaction of the
Higgs mass stability, mN ≲ 108 GeV. A similar estimate
mN ≃ 1.17 × 107 GeV; ð138Þ is obtained [70] in the case where there are three active and at
least two sterile neutrino flavors. In that case, one may use
that is, the sterile neutrino mass and, hence the the measurement of atmospheric oscillation experiments
freeze-out temperature, in our case are 2 orders of magni- for the observed active neutrino mass differences [71],
tude higher than their counterparts considered in Δm2atm ∼ 2.4 × 10−3 eV2 , and the type-I seesaw generaliza-
Refs. [36–38].26 tion of Eq. (136), giving nonzero masses to at least two
From Eq. (135), then, the corresponding Yukawa of the active neutrinos, to determine the allowed upper
coupling assumes the value jyj ≃ 4.8 × 10−5 (just a bound for mN for Higgs mass stability from experimental
factor of 5 larger than that in Ref. [38]), while from
data. Indeed, by setting mν ∼ ½Δm2νatm −1=2 in Eq. (139), one
Eq. (132) one obtains for the background field at freeze-
obtains mN ≲ ðm2H v2 ð4πÞ2 ½Δm2νatm −1=2 Þ1=3 ∼ 108 GeV. On
out [38] B0 ðT ¼ T D ≃ mN ∼ 107 GeVÞ ¼ OðkeVÞ, which
the other hand, assuming that two of the active neutrinos are
induces phenomenologically relevant leptogenesis
nearly degenerate, with the third one having a much smaller
at T ∼ 107 GeV.
mass, one may face a situation where mν ∼ Oð10−1 Þ eV,
Before closing this section, we also remark that the value
implying mN ≲ 107 GeV.
(138) is compatible with the upper bound on the sterile
As already mentioned, such naturalness bounds can be
neutrino masses required in minimal scenarios for Higgs
bypassed, if new physics, e.g., supersymmetry, exists at
mass stability (naturalness) in type-I seesaw models [70],
some scale below 107 GeV, in which case the RHN
contributions to the Higgs mass quantum corrections might
25
One needs more than one flavor for heavy neutrinos in that be canceled by, say, loops of sneutrinos, if the masses of the
case, which can be easily accommodated in the framework of latter are similar to those of the RHN. In our string-inspired
Refs. [36–38].
26
It should be noted that the freeze-out temperature could be up case, such extra contributions might well exist, but here we
to 1 order of magnitude higher than mN , due to model depend- consider minimal seesaw scenarios, which suffice for our
ence when calculating it from the equality of the total decay rate purposes.
(130) with the Hubble parameter. In such a case, one may have
q ≃ Oð100Þ or larger in the lepton asymmetry equation (131),
implying a mN ¼ Oð106 Þ or smaller, in the ballpark of the sterile
IV. MODERN ERA AND REAPPEARANCE OF THE
neutrino mass of Refs. [36–38]. Thus, the above numbers should GRAVITATIONAL ANOMALIES
be considered with a theoretical uncertainty of a couple of orders
of magnitude. The unambiguous conclusion, though, is that, in
After freeze-out, during the radiation era, the temper-
this scenario, there is phenomenologically relevant leptogenesis ature of the Universe continues to drop at a rate aðtÞ ∼ 1=T,
during the early radiation era. until the expansion of the Universe is such that the a−2 ðtÞ
045001-26
term, due to the chiral anomalies, in the solution for constant today, H ∼ H0 , which is much smaller than its
the KR axion background (111) dominates over the counterpart during inflation. Hence any gravitational
a−3 ðtÞ-scaling term. Such dominance lasts until more or anomalies would be strongly suppressed. The slow-roll
less the matter-radiation equality era, after which matter conditions for the KR axial background b̄_ are valid for
(mostly DM) begins to dominate, and this signals the dawn scaling ∼T 2 , which prompts us to conjecture a behavior
of the matter-domination epoch, which according to data today [40]
[1] ends at redshifts z ≃ 0.7, succeeded by the current
de Sitter phase. As follows from Einstein’s equations, pffiffiffiffiffiffi
during matter dominance, the scale factor behaves as b̄_ today ∼ 2ϵ0 H0 MPl ; ð141Þ
pffiffiffiffiffiffi 2=3
3 Ωm0 H 0
aðtÞ ¼ am ðtÞ ∼ t . Taking into account, as
2 in analogy to Eq. (74).27 In general, ϵ0 ≠ ϵ.
standard in cosmology, that it is only the relativistic An estimate of ϵ0 can be provided by matching the value
of b̄_
d.o.f. that contribute to the constant entropy density of
[Eq. (141)] with that of Eq. (140), upon setting
today
the Universe during its entire evolution, implies that the
am ðt0 Þ ¼ 1. We thus obtain,
matter-dominated era scale factor is inversely proportional
to temperature T, am ðtÞ ∼ T −1 , as is the case during pffiffiffi 2 2
pffiffiffiffi 3 e B ðt0 Þ
radiation dominance. ϵ0 ≃ : ð142Þ
During the matter-dominated era, then, as follows 2 4π 2 kM2Pl H0
from Eq. (109) upon imposing the continuity
assumption for the KR background and its derivatives, We proceed now to estimate the momentum scale
the a−3 ðtÞ ∼ T 3 term may be considered subdominant k of the monochromatic solution (107). This comes from
[38], with the dominant behavior being provided by the Maxwell’s equations in the presence of the chiral anoma-
a−2 ∼ T 2 chiral anomaly term [below, for convenience, lies, which for homogeneous and isotropic KR back-
we express the temperature in units of MPl , and absorb _
grounds bðtÞ read [63,65]
any T-independent proportionality constants appearing rffiffiffi
in the expression of am ðtÞ in the definition of 3 e2
Bðt0 Þ → B0 ðt0 Þ]: ∇ × BðtÞ ¼ σE − b̄_ κ BðtÞ; ð143Þ
8 4π 2
rffiffiffi
1 3 e2 1 where σ is the conductivity of charged chiral matter (we
_b̄j B2 ðt0 Þ
matter era ≃ used Ohm’s law and identified the electric current density
kMPl 2 4π 2 a2m ðtÞ
rffiffiffi as j ¼ σE). From the solution (107), one has
_ 3 e2 B02 ðt0 Þ 2
⇒ b̄jmatter era ≃ T : ð140Þ
2 4π 2 kM3Pl ∇ × BðtÞ ¼ −kBðtÞ; ð144Þ
From Eq. (140), and the above discussion, we there- and, thus, Eq. (143) becomes
fore conclude that at the late stages of the radiation era rffiffiffi
σ_ 3 e2
and during matter dominance, the presence of a chiral kBðtÞ ¼ − BðtÞ þ b̄_ κ BðtÞ: ð145Þ
anomaly implies a softer (∼T 2 ) temperature dependence k 8 4π 2
of the KR axial background, as compared to the T 3
scaling in the case of Ref. [38], where chiral anomalies The reader should bear in mind that the classical KR
were ignored. In our case, any such T 3 -scaling con- background b̄_ plays a role analogous to the chiral chemical
tribution to this background is subdominant, as follows
by continuity requirements at the interface between 27
We do not discuss here the behavior (vs the cosmic time) of
the end of radiation- and beginning of matter-domina- the gravitational anomaly during the entirety of the late de Sitter
tion eras. era. In fact, the gravitational anomaly in the current era most
During the current epoch, where matter has started to likely will not be constant, and thus will be washed out at the end
fade away, and a cosmological-constant-like (de Sitter) of the new inflationary period, which however cannot be
predicted, as the microscopic string theory dynamics leading
phase, seems, according to data [1], to start dominating the to this era is not known. Given the strongly suppressed Hubble
(accelerated) expansion of the Universe, the presence of parameter today H0 as compared to its counterpart H I during the
late-epoch gravitational waves would lead once more, inflationary period, H0 ≃ 10−55 HI , in order to ensure a constant
following the reasoning of Sec. II B, to the resurfacing gravitational anomaly à la Eqs. (69)–(70), one would need
of gravitational anomalies of the type (59) and (61); these enormously (and unnaturally) trans-Planckian cutoff values for
the momentum modes, although it must be said that, since the
can no longer be canceled by the diluted chiral matter. detailed string dynamics is unknown, one cannot make definite
However, now, the approximately constant Hubble param- statements on the subject. Fortunately, such issues do not affect
eter of the current-era de Sitter phase equals the Hubble our current study.
045001-27
potential −μ5 [63,65,72].28 However, for us, in contrast to that is, ϵ0 is of the order of the energy density of the
the considerations in Ref. [63], the KR axion is a fully magnetic field today ρB0 ¼ 12 B2 ðt0 Þ in units of the critical
fledged quantum field. ð0Þ
density of the universe ρc ¼ H20 M2Pl =3.
Ignoring chiral matter in our case, as it becomes On account of Eqs. (147) and (141), and the stiff
subdominant in the modern-era de Sitter phase, is equiv- equation of state (121) of the (massless) KR axion, which
alent to setting σ → 0 in Eq. (145).29 Taking into account dominates the “matter” part of the action in the de Sitter era,
Eq. (110), which has so far been used self-consistently, and we then observe that the latter field can provide a source of
utilizing Eq. (142), we obtain from Eq. (145) (stiff) DM, with a vacuum energy density of the order of the
magnetic field energy density.
pffiffiffi
pffiffiffiffi 3 e2
0
k≃ ϵ H0 2 ; ð146Þ 1
2 4π ρb ð¼ pb ÞjtodayðdeSitter-eraÞ ¼ ðb̄_ jtoday Þ2 ≃ B2 ðt0 Þ: ð148Þ
2
which, on account of Eqs. (145) and (141), leads to Unfortunately, in our low-energy effective string theory,
there is no way of estimating Bðt0 Þ from first principles. In
Bðt0 Þ2 2 ρB0 the context of the underlying string model, this in principle
ϵ0 ≃ ¼ ; ð147Þ
M2Pl H20 3 ρð0Þ
c
can be done by an appropriate choice of the ground state,
but in view of the landscape afflicting string theory, at
28
present such a task does not seem feasible. Thus, we have to
We note, however, that there are important subtle physical resort to phenomenological arguments.
differences between a bare μ5 and the temporal component of an
To this end, we first notice, that, as with the inflationary
axial-vector background (axial potential), such as b̄_ , coupled to phase, it is not the massless KR field which drives the late-
the axial fermion current. The latter, unlike μ5 , does not contribute
era de Sitter phase. There must be some other mechanism,
[73,74] to the so-called chiral magnetic effect (CME) [75], that is
the excitation of an electric current density in the presence of an by means of which an approximately constant potential U
external magnetic field, with a coefficient proportional to μ5 , appears dynamically during the late epochs of the Universe,
e2 which resembles quintessence, thus driving the later de
jCME ¼ 2π 2 μ5 B, which is an effect associated with the chiral
anomaly. Indeed, if one uses energy conservation arguments [73] Sitter era. In such a case, one may assume that the kinetic
or calculates the electric current density from first principles energy of the KR axion field K b ¼ 12 b_ is roughly 1 order of
using, e.g., the relativistic quantum mechanics approach magnitude smaller than U, a typical situation for other
[74] in the presence of both a chiral chemical μ5 and the
axial potential, then only the μ5 contributes to the current. In cosmological fields, such as quintessence, which would
our case, the noncontribution of the axial potential to the CME is allow the total equation of state to be approximated by that
consistent [38] with the fact that the contributions of the KR of de Sitter space-time w ≃ −1. In such a case, by
torsion [and thus the axial KR potential b̄_ ðxÞ] to the chiral identifying the two slow-roll parameters for the KR field,
anomaly can be removed by the addition of appropriate renorm- in the early and late de Sitter eras of the string Universe
alization group counterterms to the string effective action, [cf. Eq. (47) and (141)]
order by order in perturbation theory [48,49]. Such issues will
not be directly relevant for our purposes in this work, though,
which is based on Eqs. (143) and (145) and the associated ϵ ∼ ϵ0 ¼ Oð10−2 Þ ð149Þ
magnetogenesis. In this respect, the reader should notice that the
last term on the right-hand sides of Eqs. (143) and (145), which one can get the DM content in the right ballpark [1]:
has a form similar to the chiral magnetic effect, does contribute to
the magnetic field evolution, but cannot be considered as a ρm0 U Kb
contribution to the electric current, for reasons explained in Ωm0 ¼ ð0Þ
≃ ð0Þ
≃ 10 ð0Þ
≃ 10ϵ ¼ Oð0.1Þ; ð150Þ
the text. ρc ρc ρc
29
We note at this stage, that, had we kept charged chiral matter
today, and thus the conductivity σ term in Eq. (145), then the where ρm0 is the current energy density of DM in the
equation would have admitted aqffiffigrowing solution [65]
e2 3 _
universe. Above we used the fact that, according to Eqs. (73)
BðtÞ ¼ B0 expðδtÞ, with δ ¼ σk ð4π 2 8κ b̄ − kÞ > 0, for suffi- and (141), the slow-roll parameter of bðxÞ measures the ratio
ciently low k, where b̄_ is given by Eq. (141). This would have of its kinetic energy, K b ∼ ð1=2Þb_ 2 , to the critical energy
led to the well-known instabilities in the presence of a chiral density of the Universe, ρc ¼ ðMPl HÞ2 =3.
chemical potential [65], which in the approach of Ref. [63] have
been linked to the creation of a cosmological magnetic field, On account of Eq. (147), then, this also determines the
whose energy density was identified with the dark energy of the current energy density of the cosmological magnetic field,
Universe in the current era. In our approach, where chiral charged ρB0 . Moreover, we observe that the temporal component of
the KR background (129), B ¼ b̄_ M−1 in the current era
matter is not dominant in the current de Sitter era, we differ from
this interpretation, associating the KR axion with a source of dark 0 Pl
matter, as we discuss below. (141), is of order
045001-28
B0 jtoday ∼ 2.435 × 10−34 eV:

2
ð151Þ s θðxÞ
the gluons of the form g32π A Aμν
2 Gμν G ; A ¼ 1; …8 [an adjoint
SU(3) color index], where gs is the strong-interaction
We note that this is about 14 orders of magnitude larger coupling, and GAμν is the gluon field-strength tensor.
than the corresponding background found in Ref. [38], in The global Peccei-Quinn Uð1Þ symmetry is associated
the absence of chiral anomalies. with shifts aðxÞ → aðxÞ þ ϵ, where ϵ ¼ const The non-
In view of the role of the almost constant B0 ∼ b̄_ perturbative potential induced by instanton effects, which
background as a CPT- and Lorentz-symmetry-violating breaks this shift symmetry, has the schematic form
background in the effective theory, which, as mentioned
VðθÞ ¼ Λ4QCD ð1 − cosðθÞÞ, where ΛQCD is the QCD scale.
above, falls within the framework of the Standard Model
extension [68], it is imperative to check the phenomeno- Minimization of the potential fixes the strong-CP-violating
logical consistency of Eq. (151) with the current bounds of angle hθi ¼ 0. The nonperturbatively generated QCD axion
such backgrounds [76]: B0 < 10−2 eV for the temporal mass squared is m2a ¼∂ 2 V=∂a2 ja¼0 ¼Λ4QCD =f 2a and hence
component, and (the much more stringent) Bi < 10−31 GeV, ma ¼ Λ2QCD =fa .30
for the spatial components. The predicted value in In our case, as mentioned previously, the massless KR
our model (151) comfortably satisfies those bounds, even axion, which is dual in four space-time dimensions to the
if one takes into account the relative motion of our antisymmetric tensor field strength, ∂ μ b ∼ εμνρσ Hνρσ , might
laboratory frame with respect to the cosmic Robertson- be viewed [22] as the Goldstone mode of the spontaneous
Walker frame, which we take to be the CMB frame. Indeed, breaking of Lorentz symmetry induced by the constant
if the lab frame moves with a certain velocity [77] j⃗vj ≪ c background (74) or (141). However, it is possible that the
(where c is the speed of light in vacuo) with respect to shift symmetry of the KR axion is broken by some
the CMB frame, then, according to special relativity, we nonperturbative stringy effects, which are also responsible
shall also observe spatial components of Bμ in the lab frame for generating a potential for it, at least in the current
of order cosmological epoch. In such models, there might be a slow-
roll behavior of the axion, which is thus viewed as a
vi quintessence field [79], driving the current de Sitter phase
Bi ¼ γ B; i ¼ 1; 2; 3; ð152Þ
c 0 of the Universe.
There are also models that involve nontrivial interactions
where γ ∼ 1 is the Lorentz factor. As can be inferred of the KR axion with standard axions, which exist
from studies of the CMB anisotropies, a typical abundantly in string theory [44], which may thus provide
order of magnitude of the velocity of the Earth additional components of axionic DM. Below we shall
(where precision tests of the Standard Model are made) discuss such a toy model, in which the field bðxÞ acquires a
with respect to the CMB background is [1,77] potential U (and a mass) in the current epoch, and the
j⃗vj ¼ Oð390 60Þ km= sec. From Eqs. (151)–(152) expression (141) is still a consistent solution of the
then, we observe that all bounds for the Lorentz- and equations of motion. The ingredients of such a model
CPT-violating KR background Bμ , μ ¼ 0; …3, are have been considered in Ref. [43], in an attempt to propose
comfortably satisfied. alternative (beyond seesaw [69]) mechanisms for radiative
generation of right-handed Majorana neutrino masses, that
V. MASSIVE KR-AXION DARK MATTER appear, e.g., in the Lagrangian (128) and are crucial for
We would like to close our study by making some further leptogenesis. The model couples the bosonic action (102),
remarks on the nature of the KR axion as a source of DM. involving the KR axion field in the presence of a gravi-
In our approach so far, the KR axion has been treated as tational anomaly, to stringy or ordinary (including QCD)
exactly massless, not having any potential, and thus axion fields Ai ðxÞ, i ¼ 1; …n, through a kinetic mixing
respecting the shift symmetry. Axions in such conditions term [43]
are usually viewed as Goldstone bosons of a spontaneously
X
n Z
broken global (shift) symmetry (such as an accidental pffiffiffiffiffiffi
mixing ¼
Sb−a γi d4 x −g∂ μ Ai ∂ μ bðxÞ; ð153Þ
Peccei-Quinn symmetry for QCD axions [78]). If the i¼1
symmetry is broken explicitly, however, by nonperturbative
quantum (instanton) effects, as happens, for instance, in the where the (dimensionless) mixing coefficients 0 ≠ jγ i j < 1.
case of the QCD axion, then a (small) periodic axion The axions Ai are assumed to have canonically normalized
potential is developed. In this sense the axion acquires a kinetic terms and shift-symmetry-breaking nontrivial
small mass, which implies its potential role as a light DM Yukawa couplings with right-handed Majorana neutrinos,
candidate. Concretely, the QCD axion field θðxÞ ≡ aðxÞ=fa ,
where f a is the (mass-dimension-one) axion decay constant, In more precise estimates, Λ2QCD is replaced by mπ fπ, where
30
estimated phenomenologically to lie in the range [78] mπ (f π ) is the pion mass (decay constant), and the potential is
109 GeV < f a < 1012 GeV, has anomalous couplings with appropriately modified [78].
045001-29
which can be generated by nonperturbative string situation where the fields depend only on the cosmic time
instanton effects. The details of the potential of the Ai at large scales, the solution (47) is still valid despite the
fields are irrelevant for the radiative Majorana neutrino presence of the potential Uðb; A; …Þ. In that case, the
mass generation [43]. For our purposes, it suffices to equations (155) reduce to
concentrate on one such axion field AðxÞ. In general,
we assume that the A axion also couples to the (gravita-
d pffiffiffiffiffiffi _ pffiffiffiffiffiffi δUðb;A;…Þ
tional) anomaly with some dimensional coupling, which γ ½ −gðĀÞ ¼ −g ;
dt δb b¼b̄;A¼Ā
we take to be rffiffiffi
Z 2 d pffiffiffiffiffiffi ðf A −γÞα0 0
pffiffiffiffiffiffi f α0 −g K
d4 x −g A AðxÞRμνρσ R̃μνρσ 3 dt 96κ
96κ
Z pffiffiffiffiffiffi 1 δUðb;A;…Þ δUðb;A;…Þ
pffiffiffiffiffiffi f α0 ¼ −g − : ð156Þ
¼ − d4 x −g A ∂ μ AðxÞKμ ; ð154Þ γ δb δA
96κ b¼b̄;A¼Ā
where f A is a dimensionless constant, which depends Gravitational-wave perturbations contribute to the anomaly
on the microscopic details of the theory, in particular as in the inflationary period, but with a much smaller
on stringy d.o.f. circulating in the anomalous chiral Hubble parameter H0. We stress that, in a FLRW space-
fermion loop. In our current approach so far we have time, massive bðxÞ fields necessitate the presence of a
assumed for concreteness α0 ∼ κ 2 , but in realistic string-
nontrivial δUðb;A;…Þ
δb ≠ 0, and thus γ ≠ 0.
theory models one may keep the Regge slope as
In general, an approximately constant solution (141) of a
an independent parameter, to be fixed phenomenologi-
massive b axion is consistent with the above equations.
cally, and this is what we adopt for the remainder of this
Let us see this in a concrete but simple case, in which
section.
0 ≠ f A ¼ γ < 1, which implies [cf. Eq. (156)]
The equations of motion of the bðxÞ and AðxÞ fields read
(the overline above the fields, denotes classical solutions, as
per our previous notation) 1 δUðb; A; …Þ δUðb; A; …Þ
− ¼ 0: ð157Þ
rffiffiffi γ δb δA b¼b̄;A¼Ā

pffiffiffiffiffiffi α 2 α0 α α
∂ α −g ∂ b̄ − K þ γ∂ Ā
3 96κ Using Eq. (157), we observe that the first line of Eq. (156)

pffiffiffiffiffiffi δUðb; Ā; …Þ becomes
¼ −g ;
δb b¼b̄;A¼Ā
rffiffiffi _ þ Ā
̈ ¼ δUðb; A; …Þ
pffiffiffiffiffiffi α 2 f A α0 α 3HĀ : ð158Þ
∂ α −g ∂ Ā − K þ γ∂ α b̄ δA b¼b̄;A¼Ā
3 96κ

pffiffiffiffiffiffi δUðb; A; …Þ
¼ −g ; ð155Þ We remain agnostic as to the precise underlying microscopic
δA b¼b̄;A¼Ā string theory that produces the potential Uðb; AÞ through
stringy instanton effects. Therefore below we resort to
where we included a potential Uðb; A; …Þ, assumed to be phenomenological plausibility arguments. For concrete-
generated in the late cosmological eras, which explicitly ness, we assume that the axion A field induces the late
breaks the shift symmetry of the axions. Above we ignored de Sitter phase through a nonperturbatively generated
fermion and gauge anomaly contributions, as we assume (periodic) potential of a form used in inflationary scenarios
that in the current de Sitter era, fermion matter and radiation [79], which can be embedded in concrete string/brane
are not dominant, while only A- and b-axion DM theory models:
dominate.
We do not discuss here the details of the generation of the
b A
potential Uðb; A; …Þ, apart from noting that a cosmologi- Uðb; AÞ ¼ c0 M 4Pl 4
þ M1 1 − cos −
Mb MA
cal-constant-type dark energy contribution is included for
phenomenological reasons. One may use quintessence-like þ ; c0 > 0; ð159Þ
potentials, of the form used for axion inflation [79], which
contain mass terms for the bðxÞ field, so that the latter can where Mi > 0, i ¼ 1; A; b are appropriate mass scales,
play the role of an ordinary massive axionic DM compo- to be fixed phenomenologically. The term c0 M4Pl > 0 acts
nent. The important point is that, in the presence of an axion as a (positive) cosmological constant term in the current
kinetic mixing parameter γ ≠ 0 [Eq. (153)], within the era, under the slow-roll condition for the axion fields,
context of a homogeneous and isotropic cosmological which are assumed to be weak in the current epoch (see
045001-30
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
discussion below).31 The … in Eq. (159) indicates terms M4 2 M4
1 þ 9H2
−x − 2 1 2 þ 9H2 þx −
involving shift-symmetry-breaking couplings with other 3H
−2M x M M 4M M2 M 2 4M2
AðxÞ ¼ e Pl ðC̃1 e A Pl Pl
þ C̃2 e A Pl Pl
Þ;
fields, e.g., the aforementioned chiral Yukawa coupling
with right-handed fermions, ybðxÞψ̄ cR ψ R [43]. At late x ¼ tMPl ; ð163Þ
epochs, like the current one and beyond, where the
Universe enters a de Sitter phase again, we assume that where the constants C̃i ; i ¼ 1, 2, are determined by
such fermionic matter is completely diluted, or equivalently boundary conditions. In the current era, H ¼ H0 . Then,
that the corresponding Yukawa couplings (that are in general due to the smallness of H0 , we may assume for concrete-
also temperature dependent) are negligible. Hence we ignore ness that the arguments of the square roots in the exponents
them for the purposes of our subsequent discussion. on the right-hand side of Eq. (163) are negative. Upon
The reader should note that the potential (159) is imposing suitable boundary conditions, then, we arrive at a
characterized by nondiagonal mass terms for the b and dumped oscillatory solution with (increasing) cosmic time,
A fields, with the corresponding mass eigenstates obtained familiar from massive axion DM cases,
by diagonalization. The massive nature of the axions b and sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
A, then, allows them to play the role of multicomponent M41 9H2
−3H
DM in the current era. AðtÞ ¼ A0 e 2 t sin t − ;
M2A 4
The condition (157) is satisfied for the potential (159),
provided M41 9H2
A 0 ≪ MA ; > ; ð164Þ

4
M2A 4
M1 b A M1 4 b A
sin − ¼− sin −
γMb Mb MA MA Mb MA with the quantity
M
⇒ Mb ¼ − A ; ð161Þ M41 9H2
γ m2A ≡ − ð165Þ
M2A 4
where, for consistency with the condition Mi > 0; i ¼
A; b, we should take γ < 0 (the reader is reminded that playing the role of an effective axion mass squared in an
jγj < 1, but it can have either sign [43]). expanding Universe [above we kept H general, since the
We shall look for self-consistent solutions of Eq. (161) in expression (164) is valid beyond the current era]. The
which A=MA ≪ 1, to satisfy the weak-field requirement. condition A0 ≪ MA guarantees weak fields.
We shall also assume that M1 ≪ MA . Taking the kinetic Slow-roll conditions for both axions A and b, which in
mixing parameter 0 ≠ jγj ≪ 1, for concreteness, from this model behave as massive DM fields in the modern era,
Eq. (161) we observe that Mb ≫ MA , so that can thus be arranged by suitable choices of the parameters.
Eq. (158) can be approximately written, to leading order The order of Ā _ and b_ today is bounded from above
in small quantities, as by current cosmological observations [1]. Without
loss of generality, and assuming that the axions constitute
H 0 M4 the dominant form of DM today, one may assume
A00 þ 3 A ≃ − 2 1 2 A; ð162Þ
M Pl MA MPl [cf. Eq. (141)]
pffiffiffiffiffiffi
_
jAj _ 0
where the prime denotes differentiation with respect to the today ∼ jbjtoday ¼ Oð 2ϵ H 0 M Pl Þ; ð166Þ
dimensionless variable x ¼ tMPl . The general solution of
Eq. (162) is which can be easily achieved by an appropriate choice of
the parameters.
31
Alternatively, one could also consider the potential The energy density of the b − A fluid at the current
(approximately de Sitter) era is then given by

4 b A
Uðb; AÞ ¼ M1 1 − c22 cos − þ ; 0 ≠ c22 < 1; 1 _ 2 1 _ 2 γ _ _
Mb MA ρb−a
today ¼ ðĀÞ þ ðb̄Þ þ Ā b̄ þUðb̄; Ā; …Þjtoday ;
2 2 2
ð160Þ
γ ≪ 1: ð167Þ
in which the dominance of the (positive) cosmological constant
(M1 4 ð1 − c22 Þ > 0), driving the current-era de Sitter phase, arises In view of Eqs. (47) and (166), and the fact that a
from a weak-field expansion about the origin in field space cosmological constant term is present in the (slowly varying)
A ¼ b ¼ 0, corresponding to the trivial local maximum of the
potential, under the assumption of slow roll for the axion fields
potential Uðb; A; …Þ [cf. Eqs. (159) or (160)], one can
AðxÞ, bðxÞ. For the purposes of our discussion in this section, readily see that the energy density (167) in the present epoch
both potentials (159) and (160) are qualitatively equivalent. acquires a “running vacuum” form [Eq. (5)], with H 2
045001-31
contributions associated with the gravitational anomalies. contributions to the vacuum energy density, which owe
On account of the current constraints on the DM energy their existence to the gravitational anomaly. Furthermore,
density [1], and the role of both axion fields as massive DM as we have shown, in our string-inspired theoretical
components with a quintessence-like potential, we thus framework inflation can be correctly initiated and termi-
observe that Eq. (166) is consistent with the identification of nated (graceful exit) with the help of the gravitational
the slow-roll parameters of the b axion between the infla- Chern-Simons term, whose average over de Sitter space-
tionary and current eras, Eqs. (73) and (141) respectively, time also induces an additional, higher-order, power ∼H4
ϵ0 ∼ ϵ ¼ Oð10−2 Þ, as assumed in our model, following the contribution to the vacuum energy density. This higher-
argumentation leading to Eq. (150). order term triggers inflation within the context of the RVM,
This completes our discussion. We stress once more that, as has been proven in detail in the literature [8–13]. It
unfortunately, at present, the above analysis provides only follows that the entire history of the universe can be
plausibility arguments, not a concrete mechanism for mass described in an effective RVM language upon starting
generation for the KR axion, due to the lack of knowledge from the fundamental massless bosonic gravitational multi-
of the underlying microscopic string/brane model that plet of a generic string theory. We believe that this is an
could generate the (nonperturbative) potential Uðb; AÞ. interesting and remarkable result of our work, which, to the
Nonetheless, we believe that the arguments are sufficiently best of our knowledge, was never put forward in the
interesting to foster further research in this direction. The literature prior to the present work. Thanks to this result,
fact that our model promotes axionic DM as the dominant the effective language of the RVM can be used in a very
species of DM in the Universe, makes it relevant for current practical way to compute the main traits of the cosmic
DM studies, in particular in models in which the effective evolution starting from inflation and going through the
DM mass (165) is small, so that the respective DM is standard radiation- and matter-dominated epochs until the
ultralight. Such ultralight DM currently constitutes the late-time universe, i.e., the incipient DE epoch around our
subject of intense research, proposing, for instance, the time, and finally into the future.
use of precision atomic or laser interferometric devices or Because of the anomalous coupling of the KR axion to
other quantum sensors, to falsify particle physics models gravitational anomalies, the field remains undiluted at the
involving scalar or pseudoscalar (axion) DM particles with end of inflation. During the radiation/matter eras, chiral
masses smaller than 10−21 eV [80]. fermionic matter generated at the end of inflation cancels
the gravitational anomalies, thus restoring diffeomorphism
VI. CONCLUSIONS invariance in the radiation/matter quantum field theory, as
In this article we have provided a string-inspired theo- required for consistency. We have found that, as the
retical framework in which, during the early phase of the Universe passes from inflation to the radiation-dominated
Universe, there are important contributions to the vacuum epoch, the presence of the undiluted CP- and (sponta-
energy density which are related to the CP-violating neously) CPT-violating KR axion background, may lead to
gravitational anomalies of a primordial space-time of baryogenesis via leptogenesis, in models involving heavy
string theory. The latter are induced by primordial gravi- right-handed (sterile) neutrinos. The lepton asymmetry is
tational waves during the inflationary era, in the presence of generated by CP (and CPT)-violating decays of the sterile
Lorentz- and CPT-violating backgrounds of the KR axion neutrinos into Standard Model particles in the presence of
field of the massless bosonic string multiplet. The KR field the KR background. Baryon-lepton-number-conserving
itself, though, does not cause or drive inflation, which is sphaleron processes in the Standard Model sector of the
due to other independent mechanisms. theory can then communicate the lepton asymmetry to
During the primordial inflationary era, we assume that baryons, thus leading to baryogenesis. Therefore, the
only (stringy) gravitational d.o.f. are present. Hence, the aforementioned process could provide an efficient way
gravitational anomalies, whose presence in general would to understand the underlying physical mechanism for the
cause diffeomorphism-invariance breaking in the quantum dominance of matter over antimatter in the early Universe.
theory, do not constitute any inconsistency, as would be the Moreover, during the radiation/matter dominance, uncom-
case if matter were present, since the anomalies describe the pensated gauge chiral anomalies of the fermionic-matter
exchange of energy solely among (quantum) gravitational axial (chiral) currents, also lead to H2 ðtÞ “running vac-
d.o.f. Moreover, there is a second-rank modified stress uum”-type contributions to the energy density of the
tensor which is conserved and describes any exchange of Universe.
energy between the KR axion field and gravity. The stress In the late universe such running vacuum contributions
tensor of this KR axion alone, which would be the “matter” involve an additive constant term, which was neglected in
stress tensor if anomalies were absent, is not conserved in the early universe, and hence the effective or “running”
their presence. It is important to mention that the infla- cosmological term within the RVM is of the form
tionary epoch can be described using the formalism of an ΛðHÞ ¼ c0 þ νH2 , where the value of c0 is close (but
effective “running vacuum” model with H2 -type not exactly equal) to the cosmological constant term of the
045001-32
ΛCDM model, and νH2 (with jνj ≪ 1Þ is the running part or chiral anomalies lead to mildly running dark energy, as a
of the DE density, a characteristic feature of the model. The smoking gun evidence of their presence!
RVM is, therefore, finally testable in a very concrete way. It So, paraphrasing the famous quote by Carl Sagan [82],
provides a specific mechanism for inflation, which is the thesis of this article is that we may well be anomalously
different from the conventional one based on the inflaton made of star stuff!
field [8–13], but also furnishes a mildly varying vacuum Affaire à suivre...
contribution which surfaces in the late universe and
can be perceived as a form of dynamical dark energy. ACKNOWLEDGMENTS
Such a form of DE leads to an overall improvement of the We would like to thank the anonymous referee for
fit of the cosmological observations as compared to the case constructive and insightful comments, that helped in
of a rigid Λ term [14–17]. In addition, that dynamical improving the presentation of our results. N. E. M. wishes
component of the DE can help alleviate some of the to acknowledge discussions with Diego Blas and Chris
tensions that exist in the ΛCDM model concerning σ 8 McCabe on axion DM models and related searches. S. B.
[15] and H0 [31]. acknowledges support from the Research Center for
During the current de Sitter era, the dilution of any Astronomy of the Academy of Athens in the context of
matter, and the dominance again of the stringy gravitational the program “Tracing the Cosmic Acceleration.” The work
d.o.f., including the KR axion, leads, through late-epoch of N. E. M. is supported in part by the UK Science and
gravitational waves, to the resurfacing of gravitational Technology Facilities research Council (STFC) under the
anomalies. We have also discussed how the KR field in research Grant No. ST/P000258/1. The work of J. S. has
the present era can act as a source for dark matter in models been partially supported by Projects No. FPA2016-76005-
involving large-scale cosmic magnetic fields, generated by C2-1-P (MINECO), 2017-SGR-929 (Generalitat de
the chiral anomalies. The magnetic energy density con- Catalunya) and MDM-2014-0369 Institute of Cosmos
tributes to the late-era energy budget of the Universe, with Sciences of the Universitat de Barcelona (ICCUB). This
terms of RVM type, scaling as H20 . Moreover, there are work is also partially supported by the COST Association
Action CA18108 “Quantum Gravity Phenomenology in the
scenarios in which the KR field mixes with other axion
Multimessenger Approach (QG-MM).” N. E. M. acknowl-
fields, which are abundant, e.g., in string models, thus
edges a scientific associateship (“Doctor Vinculado”) at
providing models for multicomponent DM. IFIC-CSIC-Valencia University, Valencia, Spain.
Before closing, we make a last but rather important
remark. Since our effective-field-theoretic running-vacuum
APPENDIX: ARBITRARY STRING MASS SCALE
model of quantum gravity, upon which we based our
studies here, is inspired by string theory, it would be In this appendix, we demonstrate how in the general
interesting to discuss it in the context of the recent case, where the string scale Ms ≠ MPl , one can avoid trans-
conjectures on the incompatibility of de Sitter vacua Planckian values for the UV cutoff μ by appropriately
(characterized by a rigid positive cosmological constant) constraining the range of M s .
with the “swampland criteria,” and in general of how one Indeed, this follows from the fact that, in such a case, the
can couple quantum field theories to quantum gravity condition (70) for an approximately constant K0 during
models, especially in view of our trans-Planckian regime inflation is replaced by
of modes entering Eq. (71) [81]. We leave this interesting
−5 H 2 μ 4
topic for future works, as we did not discuss here micro- A ¼ 1 − 1.95 × 10 ≃0
scopic string theory models leading to inflation. We MPl Ms

remark, though, that the dynamical nature of the running μ M Pl 1=2
⇒ ≃ 15 ; ðA1Þ
vacuum leads to deviations of our model from the standard Ms H
ΛCDM model, as far as the nature of the vacuum energy is
where the ≃ in the above relations are to be interpreted as
concerned, which is dynamical in our case; in this respect,
within an error of order at most 1%. Indeed, an approx-
the compatibility of some of the models falling within our
imately constant K0 in Eq. (68) is guaranteed provided that
framework with the “swampland criteria” is to be expected.
at the end of the inflationary period its value is diminished
To summarize our findings, the (gravitational) anomaly
by no more than an order of magnitude, that is
played an important dual role for our existence: first, it
induced a nondiluted axion background of DM at the end of K0end ðtend Þ ≃ ðe−1 − e−2 ÞK0begin ðtðη ¼ H−1 ÞÞ: ðA2Þ
inflation into the radiation epoch, which itself induces
leptogenesis; and, second, it fostered the subsequent gen- Taking in to account that, in units of cosmic Robertson-
eration of chiral matter from the decay of the running Walker time t, the end of inflation occurs for Htend ∼ N ,
vacuum, thus canceling the unbalanced gravitational where N , the number of e-foldings, is expected from the
anomaly and restoring general covariance in our data [1] to be of order N ¼ Oð60–70Þ, we thus observe
Universe. As demonstrated in our work, the gravitational from Eq. (A2) that
045001-33
0 ≲ A ≲ ξð3N Þ−1 ∼ ξð0.0048 − 0.0056Þ; ξ ¼ 1 − 2; ðA3Þ which, on account of Eq. (A5) implies
suffices for our purposes, which leads to the aforemen-

μ ≫ 2.61 × ð10−3 –10−2 ÞM Pl : ðA7Þ
tioned uncertainty of at most 1% in the value of μ in
Eq. (A1),
This, in turn, leads to the observation that the cutoff scale μ
μ ξ 1=4 MPl 1=2
≃ 15 1 − can be at least of order M Pl. Thus, by allowing M s ≠ M Pl ,
Ms 3N H we can in principle avoid a trans-Planckian cutoff μ, since

M Pl 1=2 we may set μ ∼ M Pl which is a quite natural order of
≃ ð0.998–0.999Þ × 15 : ðA4Þ
H magnitude for the UV completion of the low-energy
effective theory. In such a case, Eq. (A6) implies the
If one insists on phenomenologically acceptable ranges of following range of the minimal allowed order of magnitude
H ≪ M Pl , e.g., Eq. (64), then one obtains
of the string scale: Ms ≳ 10−3 M Pl . Saturating from above
μ ∼ 103 M s ; ðA5Þ Ms ≲ MPl we thus obtain the range for the string scale
which replaces Eq. (71). Then, upon combining Eqs. (61)

and (62), we see that a sufficient condition to guarantee the M Pl ≳ M s ≳ 10−3 MPl ; ðA8Þ
smallness of jΘ ≪ 1 is
in order to guarantee the Lorentz-violating solution (72) for
H=Ms ≪ 3.83; ðA6Þ the KR background.
[1] N. Aghanim et al. (Planck Collaboration), arXiv:1807. [16] J. Solà, A. Gómez-Valent, and J. de Cruz Pérez, Phys. Lett.
06209. B 774, 317 (2017); Astrophys. J. 836, 43 (2017); 811, L14
[2] S. Weinberg, Rev. Mod. Phys. 61, 1 (1989); P. J. Peebles and (2015).
B. Ratra, Rev. Mod. Phys. 75, 559 (2003); T. Padmanabhan, [17] A. Gómez-Valent, J. Solà, and S. Basilakos, J. Cosmol.
Phys. Rep. 380, 235 (2003). Astropart. Phys. 01 (2015) 004.
[3] I. L. Shapiro and J. Solà, J. High Energy Phys. 02 (2002) [18] J. Grande, J. Solà, S. Basilakos, and M. Plionis, J. Cosmol.
006; Phys. Lett. B 475, 236 (2000); Nucl. Phys. B, Proc. Astropart. Phys. 08 (2011) 007.
Suppl. 127, 71 (2004); Proc. Sci., AHEP2003 (2003) 013. [19] S. Basilakos, M. Plionis, and J. Solà, Phys. Rev. D 80,
[4] J. Solà, J. Phys. A 41, 164066 (2008). 083511 (2009).
[5] I. L. Shapiro and J. Solà, Phys. Lett. B 682, 105 (2009). [20] C. España-Bonet et al., J. Cosmol. Astropart. Phys. 02
[6] J. Solà, J. Phys. Conf. Ser. 453, 012015 (2013); AIP Conf. (2004) 006; Phys. Lett. B 574, 149 (2003).
Proc. 1606, 19 (2014); J. Phys. Conf. Ser. 283, 012033 [21] J. R. Ellis, N. E. Mavromatos, and D. V. Nanopoulos, Gen.
(2011); Proc. Sci., AHEP2003 (2013) 013. Relativ. Gravit. 32, 943 (2000).
[7] J. Solà and A. Gómez-Valent, Int. J. Mod. Phys. D 24, [22] I. Antoniadis, C. Bachas, J. R. Ellis, and D. V. Nanopoulos,
1541003 (2015). Nucl. Phys. B328, 117 (1989).
[8] S. Basilakos, J. A. S. Lima, and J. Solà, Mon. Not. R. [23] J. R. Ellis, N. E. Mavromatos, and D. V. Nanopoulos, Phys.
Astron. Soc. 431, 923 (2013). Lett. B 293, 37 (1992).
[9] E. L. D. Perico, J. A. S. Lima, S. Basilakos, and J. Solà, [24] For a very partial list of representative works in this topic
Phys. Rev. D 88, 063531 (2013). see: H. Terazawa, Phys. Lett. 101B, 43 (1981); M. Ozer and
[10] S. Basilakos, J. A. S. Lima, and J. Solà, Int. J. Mod. Phys. D M. O. Taha, Nucl. Phys. B287, 776 (1987); O. Bertolami,
22, 1342008 (2013); 23, 1442011 (2014). Nuovo Cimento B 93, 36 (1986);K. Freese, F. C. Adams,
[11] S. Basilakos, J. A. S. Lima, and J. Solà, Gen. Relativ. Gravit. J. A. Frieman, and E. Mottola, Nucl. Phys. B287, 797 (1987);
47, 40 (2015). J. C. Carvalho, J. A. S. Lima, and I. Waga, Phys. Rev. D 46,
[12] J. Solà, Int. J. Mod. Phys. D 24, 1544027 (2015). 2404 (1992); J. M. Overduin and F. I. Cooperstock, Phys.
[13] S. Basilakos, J. A. S. Lima, and J. Solà, Eur. Phys. J. C 76, Rev. D 58, 043506 (1998); For reviews see also: J. P. Uzan,
228 (2016). Living Rev. Relativity 14, 2 (2011), and references therein; T.
[14] J. Solà, J. de Cruz Pérez, and A. Gómez-Valent, Europhys. Chiba, Prog. Theor. Phys. 126, 993 (2011).
Lett. 121, 39001 (2018); Mon. Not. R. Astron. Soc. 478, [25] I. L. Shapiro, J. Solà, and H. Štefančić, J. Cosmol. Astropart.
4357 (2018). Phys. 01 (2005) 012.
[15] A. Gómez-Valent and J. Solà, Mon. Not. R. Astron. Soc. [26] S. Basilakos, N. E. Mavromatos, and J. Solà, Universe 2, 14
478, 126 (2018); Europhys. Lett. 120, 39001 (2017). (2016).
045001-34
[27] T. Appelquist and J. Carazzone, Phys. Rev. D 11, 2856 [51] P. Kanti, N. E. Mavromatos, J. Rizos, K. Tamvakis, and E.
(1975). Winstanley, Phys. Rev. D 54, 5049 (1996).
[28] H. Fritzsch and J. Solà, Classical Quantum Gravity 29, [52] S. H. S. Alexander, M. E. Peskin, and M. M. Sheikh-Jabbari,
215002 (2012). Phys. Rev. Lett. 96, 081301 (2006); eConf C0605151, 0022
[29] A. Babić, B. Guberina, R. Horvat, and H. Štefančić, Phys. (2006).
Rev. D 71, 124041 (2005); 65, 085002 (2002). [53] K. S. Stelle, Gen. Relativ. Gravit. 9, 353 (1978); B. Whitt,
[30] A. G. Riess, S. Casertano, W. Yuan, L. M. Macri, and D. Phys. Lett. 145B, 176 (1984).
Scolnic, Astrophys. J. 876, 85 (2019); 853, 126 (2018); 826, [54] A. A. Starobinsky, Phys. Lett. 91B, 99 (1980); 91B, 99
56 (2016). (1980); Adv. Ser. Astrophys. Cosmol. 3, 130 (1987).
[31] M. Rezaei, M. Malekjani, and J. Solà, Phys. Rev. D 100, [55] For a very partial but representative list of references
023539 (2019). discussing the trans-Planckian problem in inflationary
[32] M. B. Green, J. H. Schwarz, and E. Witten, Superstring cosmology from various points of view, see: J. Martin
Theory. Vols. 1: Introduction, Cambridge Monographs On and R. H. Brandenberger, Phys. Rev. D 63, 123501 (2001); R.
Mathematical Physics (Univ. Pr., Cambridge, UK, 1987), H. Brandenberger, arXiv:hep-th/0210186; U. H. Danielsson,
P. 469; Superstring Theory. Vol. 2: Loop Amplitudes, Phys. Rev. D 66, 023511 (2002); A. A. Starobinsky, Pis’ma
Anomalies And Phenomenology, Cambridge Monographs Zh. Eksp. Teor. Fiz. 73, 415 (2001) [JETP Lett. 73, 371
On Mathematical Physics (Univ. Pr., Cambridge, Uk, 1987), (2001)]; J. Martin and R. H. Brandenberger, arXiv:astro-ph/
P. 596. 0012031.
[33] D. J. Gross and J. H. Sloan, Nucl. Phys. B291, 41 (1987); R. [56] D. H. Lyth, C. Quimbay, and Y. Rodriguez, J. High Energy
R. Metsaev and A. A. Tseytlin, Nucl. Phys. B293, 385 Phys. 03 (2005) 016.
(1987); M. C. Bento and N. E. Mavromatos, Phys. Lett. B [57] W. Fischler and S. Paban, J. High Energy Phys. 10 (2007)
190, 105 (1987). 066.
[34] M. J. Duncan, N. Kaloper, and K. A. Olive, Nucl. Phys. [58] H. Collins and R. Holman, Phys. Rev. D 77, 105016
B387, 215 (1992). (2008).
[35] N. E. Mavromatos and S. Sarkar, Eur. Phys. J. C 73, 2359 [59] J. Alexandre, N. Houston, and N. E. Mavromatos, Phys.
(2013). Rev. D 89, 027703 (2014); Int. J. Mod. Phys. D 24, 1541004
[36] M. de Cesare, N. E. Mavromatos, and S. Sarkar, Eur. Phys. (2015).
J. C 75, 514 (2015). [60] S. Basilakos, N. E. Mavromatos, and J. Solà, J. Cosmol.
[37] T. Bossingham, N. E. Mavromatos, and S. Sarkar, Eur. Phys. Astropart. Phys. 12 (2019) 025.
J. C 78, 113 (2018). [61] G. de Berredo-Peixoto, L. Freidel, I. L. Shapiro, and C. A.
[38] T. Bossingham, N. E. Mavromatos, and S. Sarkar, Eur. Phys. de Souza, J. Cosmol. Astropart. Phys. 06 (2012) 017.
J. C 79, 50 (2019); see also: N. E. Mavromatos and S. [62] S. L. Adler, Phys. Rev. 177, 2426 (1969); J. S. Bell and R.
Sarkar, Universe 5, 5 (2018). Jackiw, Nuovo Cimento A 60, 47 (1969).
[39] V. A. Kuzmin, V. A. Rubakov, and M. E. Shaposhnikov, [63] A. R. Zhitnitsky, Phys. Rev. D 99, 103518 (2019), and
Phys. Lett. 155B, 36 (1985). references therein.
[40] S. Basilakos, N. E. Mavromatos, and J. Solà, Int. J. Mod. [64] B. Mukhopadhyay, Classical Quantum Gravity 24, 1433
Phys. D 28, 1944002 (2019); (essay awarded Honorable (2007).
Mention by the Gravity Research Foundation in the 2019 [65] A. Boyarsky, J. Frohlich, and O. Ruchayskiy, Phys. Rev.
Essay Competition); see also, Proc. Sci., CORFU2018 Lett. 108, 031301 (2012); Phys. Rev. D 92, 043004
(2019) 044 [arXiv:1905.05685]. (2015).
[41] R. Jackiw and S. Y. Pi, Phys. Rev. D 68, 104012 (2003). [66] M. Safarzadeh and A. Loeb, Astrophys. J. 877, L27 (2019);
[42] Y. B. Zel’dovich, Mon. Not. R. Astron. Soc. 160, 1 (1972); R. Cota, J. Barranco, D. Delepine, and S. Khalil, J. Phys.
Sov. Phys. JETP 14, 1143 (1962); P. H. Chavanis, Phys. 468, 012003 (2013); Proc. Sci., DSU2012 (2012) 026.
Rev. D 92, 103004 (2015). [67] G. W. Gibbons and S. W. Hawking, Phys. Rev. D 15, 2738
[43] N. E. Mavromatos and A. Pilaftsis, Phys. Rev. D 86, 124038 (1977); see also, in a more general context of Euclidean
(2012). quantum gravity: Euclidean Quantum Gravity (World
[44] A. Arvanitaki, S. Dimopoulos, S. Dubovsky, N. Kaloper, Scientific, Singapore, 1993).
and J. March-Russell, Phys. Rev. D 81, 123530 (2010). [68] D. Colladay and V. A. Kostelecky, Phys. Rev. D 58, 116002
[45] G. B. Zhao et al., Nat. Astron. 1, 627 (2017). (1998).
[46] N. E. Mavromatos and J. Rizos, Int. J. Mod. Phys. A 18, 57 [69] P. Minkowski, Phys. Lett. B 67, 421 (1977); M. Gell-Mann,
(2003); Phys. Rev. D 62, 124004 (2000). P. Ramond, and R. Slansky, in Supergravity, edited by D. Z.
[47] F. W. Hehl, P. Von Der Heyde, G. D. Kerlick, and J. M. Freedman and P. van Nieuwenhuizen (North-Holland,
Nester, Rev. Mod. Phys. 48, 393 (1976); T. Watanabe, Amsterdam, 1979); T. Yanagida, in Proceedings of the
arXiv:0902.1392; B. P. Dolan, Classical Quantum Gravity Workshop on the Unified Theory and the Baryon Number in
27, 095010 (2010); 27, 249801(E) (2010), and references the Universe, Tsukuba, Japan, edited by O. Sawada and A.
therein. Sugamoto (1979), http://inspirehep.net/record/19833; R. N.
[48] C. M. Hull, Phys. Lett. 167B, 51 (1986). Mohapatra and G. Senjanovic, Phys. Rev. Lett. 44, 912
[49] N. E. Mavromatos, J. Phys. A 21, 2279 (1988). (1980); J. Schechter and J. W. F. Valle, Phys. Rev. D 22,
[50] See for instance: R. Bluhm, Phys. Rev. D 91, 065034 2227 (1980); G. Lazarides, Q. Shafi, and C. Wetterich, Nucl.
(2015). Phys. B181, 287 (1981).
045001-35
[70] F. Vissani, Phys. Rev. D 57, 7027 (1998); for recent related [78] J. E. Kim and G. Carosi, Rev. Mod. Phys. 82, 557
work see, for instance: G. Bambhaniya, P. S. Bhupal Dev, (2010).
S. Goswami, S. Khan, and W. Rodejohann, Phys. Rev. D 95, [79] M. Berg, E. Pajer, and S. Sjors, Phys. Rev. D 81, 103535
095016 (2017). (2010). For a review of axion inflation see: E. Pajer and
[71] M. Tanabashi et al. (Particle Data Group), Phys. Rev. D 98, M. Peloso, Classical Quantum Gravity 30, 214002 (2013),
030001 (2018). and references therein.
[72] For magnetogenesis and other interesting physical applica- [80] See, for instance: L. Hui, J. P. Ostriker, S. Tremaine, and E.
tions of the connection of KR fields to torsion and chiral Witten, Phys. Rev. D 95, 043541 (2017); Y. V. Stadnik and
chemical potential, as well as other cosmic consequences, V. V. Flambaum, Phys. Rev. D 89, 043522 (2014); Phys.
like non-trivial ‘optical’ activity in the Universe, see: Rev. Lett. 115, 201301 (2015); A. Aoki and J. Soda, Int. J.
L. C. G. de Andrade, Classical Quantum Gravity 33, Mod. Phys. D 26, 1750063 (2017); C. Abel et al., Phys.
025006 (2016); Eur. Phys. J. C 78, 530 (2018); P. Majumdar Rev. X 7, 041034 (2017); P. W. Graham, D. E. Kaplan, J.
and S. SenGupta, Classical Quantum Gravity 16, L89 Mardon, S. Rajendran, W. A. Terrano, L. Trahms, and T.
(1999); S. Kar, P. Majumdar, S. SenGupta, and S. Sur, Wilkason, Phys. Rev. D 97, 055006 (2018); Z. Ahmed et al.,
Classical Quantum Gravity 19, 677 (2002). arXiv:1803.11306, and references therein; For a compre-
[73] D. B. Kaplan, S. Reddy, and S. Sen, Phys. Rev. D 96, hensive review on some detection methods see also: Y. V.
016008 (2017). Stadnik, Manifestations of Dark Matter and Variations of
[74] M. Dvornikov, Phys. Rev. D 98, 036016 (2018). the Fundamental Constants in Atoms and Astrophysical
[75] H. B. Nielsen and M. Ninomiya, Phys. Lett. 130B, 389 Phenomena (Springer Thesis, 2017), ISBN: 978-3-319-
(1983); K. Fukushima, D. E. Kharzeev, and H. J. Warringa, 63417-3.
Phys. Rev. D 78, 074033 (2008). [81] P. Agrawal, G. Obied, and C. Vafa, arXiv:1906.08261; H.
[76] V. A. Kostelecky and N. Russell, Rev. Mod. Phys. 83, 11 Ooguri, E. Palti, G. Shiu, and C. Vafa, Phys. Lett. B 788,
(2011); see also: N. E. Mavromatos, Hyperfine Interact. 180 (2019); S. K. Garg and C. Krishnan, J. High Energy
228, 7 (2014); V. A. Kostelecky and M. Mewes, Phys. Rev. Phys. 11 (2019) 075; G. Obied, H. Ooguri, L. Spodyneiko,
Lett. 99, 011601 (2007); M. Das, S. Mohanty, and A. R. and C. Vafa, arXiv:1806.08362; P. Agrawal, G. Obied, P. J.
Prasanna, Int. J. Mod. Phys. D 22, 1350011 (2013). Steinhardt, and C. Vafa, Phys. Lett. B 784, 271 (2018); H.
[77] See, e.g.: E. K. Conklin, Nature (London) 222, 971 (1969); Ooguri and C. Vafa, Nucl. Phys. B766, 21 (2007).
G. F. Smoot, M. V. Gorenstein, and R. A. Muller, Phys. Rev. [82] https://www.goodreads.com/author/quotes/10538.Carl-
Lett. 39, 898 (1977). Sagan.
045001-36

PhysRevD 101 045001

Uploaded by

Copyright:

Available Formats

PhysRevD 101 045001

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

PhysRevD 101 045001

Uploaded by

Copyright:

Available Formats

PHYSICAL REVIEW D 101, 045001 (2020)

Gravitational and chiral anomalies in the running vacuum universe

(Received 22 July 2019; accepted 8 January 2020; published 5 February 2020)

I. INTRODUCTION AND MOTIVATION: RUNNING

2470-0010=2020=101(4)=045001(36) 045001-1 Published by the American Physical Society

because of the dimensionality of ρΛ , the first allowed term _ ¼ a0 þ a1 H

To compute the metric variation of Eq. (30), we take into 1

R η 0 the inflation period, Eq. (67) can be integrated over

rffiffiffi rffiffiffi rffiffiffi 2

1 pffiffiffiffiffiffi e2 fields, corresponding to a single mode of momentum

By comparing Eq. (116) with Eq. (118), the reader can d F d

rffiffiffi Indeed, as discussed in Refs. [36–38], the presence of the

where s is the entropy density of the Universe, T D denotes

MPl that is, assuming no new physics (such as supersymmetry)

I ∼ 103 ð137Þ mN ≲ ðm2H v2 ð4πÞ2 m−1

B0 jtoday ∼ 2.435 × 10−34 eV:

suffices for our purposes, which leads to the aforemen-

which replaces Eq. (71). Then, upon combining Eqs. (61)

You might also like

PhysRevD 101 045001

Uploaded by

Copyright:

Available Formats

PhysRevD 101 045001

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

PhysRevD 101 045001

Uploaded by

Copyright:

Available Formats

PHYSICAL REVIEW D 101, 045001 (2020)

Gravitational and chiral anomalies in the running vacuum universe

(Received 22 July 2019; accepted 8 January 2020; published 5 February 2020)

I. INTRODUCTION AND MOTIVATION: RUNNING

2470-0010=2020=101(4)=045001(36) 045001-1 Published by the American Physical Society

because of the dimensionality of ρΛ , the first allowed term _ ¼ a0 þ a1 H

To compute the metric variation of Eq. (30), we take into 1

R η 0 the inflation period, Eq. (67) can be integrated over

rffiffiffi rffiffiffi  rffiffiffi 2

1 pffiffiffiffiffiffi e2 fields, corresponding to a single mode of momentum

By comparing Eq. (116) with Eq. (118), the reader can d F d

rffiffiffi Indeed, as discussed in Refs. [36–38], the presence of the

where s is the entropy density of the Universe, T D denotes

MPl that is, assuming no new physics (such as supersymmetry)

I ∼ 103 ð137Þ mN ≲ ðm2H v2 ð4πÞ2 m−1

B0 jtoday ∼ 2.435 × 10−34 eV:

suffices for our purposes, which leads to the aforemen-

which replaces Eq. (71). Then, upon combining Eqs. (61)

You might also like

rffiffiffi rffiffiffi rffiffiffi 2