Vacuum energy and cosmological inflation

Author: Daniel Tuells Miralles

Facultat de Fı́sica, Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain.

Advisor: Dr. Joan Solà Peracaula

Abstract: We study the cosmic evolution from the perspective of the Running Vacuum Models,
in which the vacuum energy density is dynamic throughout the cosmic history. In this context, we
find that there is an initial period of inflation, in which the vacuum decays into radiation and there
is a huge entropy production. In contrast to the standard inflaton-based mechanism of inflation,
there is a smooth transition between this period and the standard ΛCDM radiation epoch. We also
test the model against the Generalised Second Law, finding it is fulfilled in the current universe but
not in the early stages. This is a welcome feature, however, as it may provide a possible solution to
both the entropy and horizon problems, in a context where the eventual thermodynamic equilibrium
of the universe is not jeopardised.

I. INTRODUCTION favourable results.

The aim of this work is to solve the cosmological equa-
The ΛCDM model has been the most successful cosmo- tions of the RVMs so as to find expressions for the Hubble
logical theory, due to its ability to explain the accelerated parameter as well as the matter and radiation densities,
expansion of the Universe, the Cosmic Microwave Back- both in the early and current universe; we will see that
ground (CMB), the Big Bang Nucleosynthesis (BBN) these solutions provide a smooth transition between the
and the observed abundances of elements, among others. inflation and the radiation epoch. We will also study the
Nevertheless, the model is not exempt of some severe lim- model from a thermodynamic point of view, for which we
itations, as the well-known Hubble tension, for example, shall adopt the framework of the Generalised Second Law
evinces: the incompatible measurements of the Hubble (GSL). Lastly, we will analyse the ability of the model to
parameter H are hard to explain in this context, while solve the horizon and entropy problems.
a non-constant Cosmological Constant (CC), Λ, may be
able to alleviate this tension [1]. But perhaps the most
II. RUNNING VACUUM MODELS
well-known issue of the ΛCDM model is the so-called
CC problem, which arises when one tries to compare
the value for the vacuum energy density, ρΛ = Λ/8πG, We will focus our study on the type of models, de-
obtained via observations (ρ0Λ ∼ 10−47 GeV4 ) and theo- veloped in the theoretical framework of QFT in curved
retical predictions: even if we only consider the lowest space-time, in which the vacuum energy density is a
contribution from the electroweak vacuum energy, the dynamical parameter throughout the cosmic evolution,
discrepancy is already [2–4]: ρ̇Λ ̸= 0. Notice that a time-evolving gravitational con-
 EW 0  stant G = G(t) has also been discussed in the literature
ρΛ /ρΛ  = O(1055 ) , (1) (see [2, 11]), however we will restrict ourselves to the
case G = const. Using the FLRW metric one can obtain
which is perhaps the most disastrous prediction in the the following renormalisation group equation (we will be
history of science. On top of that, there are the horizon using natural units) [2]:
and entropy problems [5, 6], which will be explained in
detail later on. dρΛ 1 X
ai Mi2 H 2 + bi H 4 + O H 6 . (2)


2
= 2
In view of these issues, it has been proposed in the d ln H (4π) i
literature a type of Running Vacuum Models (RVMs),
in which the vacuum energy density is dynamic through- The motivation and derivation of this equation can be
out the entire cosmic history and runs with the Hubble found in [3]. Here the index i refers to the contributions
rate, ρΛ (H); this dynamical dependence comes from the of boson and fermion matter fields with masses Mi . It
calculation of quantum effects in QFT in curved space- can be seen [2] that the solution is of the form:
time (see [2–6] and references therein). We will try to 3

H4

2
show that these RVMs may involve the necessary theo- ρΛ (H) = c0 + νH + α 2 , (3)
8πG HI
retical ingredients to provide a solution (or at least an
alleviation) to the rest of the aforementioned problems, where the dimensionless coefficients ν, α are related to
while at the same time complying with the laws of ther- the parameters ai , bi , Mi in (2), while HI is the Hubble
modynamics. The RVMs have also been tested against rate at the scale of inflation. A more general solution
observational data from Supernovae type Ia, the CMB allows H 2n+2 terms, instead of the O(H 4 ) term, to ap-
shift parameter and Baryonic Acoustic Oscillations [7–9]; pear in (3). It leads, however, to the same conclusions
as well as Large Scale Structure formation [6–10], with we shall find, see Apendix B in [5].
Vacuum energy and cosmological inflation Daniel Tuells Miralles

The O(H 4 ) term dominates the vacuum energy evolu- where prime denotes differentiation with respect to the
tion in the early inflationary universe, when the Hubble scale factor, and we have used the chain rule d/dt =
rate is comparable to HI , while in the present universe aHd/da. Solving for both the matter and radiation den-
it is negligible due to the small value at the current time sity leads to [5]:
H0 . This allows us to set the value for the constant pa-
rameter c0 (which in this context would be nothing but ρm (a) = ρ0m a−3(1−ν) , (10)
the ”actual” cosmological constant!) in terms of the cur- ρr (a) = ρ0r a−4(1−ν) . (11)
rent vacuum energy density ρ0Λ :
For ν = 0 we recover the standard solutions, as expected.
8πG 0 Inserting (10) in (9) one finds for the matter epoch:
c0 = ρ − νH02 = H02 (Ω0Λ − ν) . (4)
3 Λ
νρ0m −3(1−ν)
In the last equality we have used the standard definitions ρΛ (a) = ρ0Λ + (a − 1) . (12)
1−ν
Ω0Λ = ρ0Λ /ρ0c and ρ0c = 3H02 /8πG. After inflation, when
H ≪ HI , the dynamics of ρΛ begin to be dominated Substituting these formulas in (6):
by the O(H 2 ) term. Here, a small ν term ensures a mild
Ω0
 
evolution of the vacuum energy, or in other words a small H 2 (a) = H02 1 + m (a−3(1−ν) − 1) , (13)
correction to the ΛCDM model. Indeed, confronting the 1−ν
model with observations one finds |ν| < −3
∼ 10 [7, 8, 11], where Ω0m = ρ0m /ρ0c and Ω0m + Ω0Λ = 1 as usual. In these
while a thermodynamic analysis leads us to conclude that last results we verify, once again, that the ν-term seems
ν > 0 [5]. Therefore, the CC problem would be nothing to act as a small dynamical correction to the cosmological
but a consequence of living in a very low energy uni- constant.
verse where the vacuum energy density is almost equal
to 3c0 /8πG, which, following the ΛCDM model, leads us
to believe it has remained constant throughout the entire B. Early universe
cosmic history!
We shall now derive some useful relations to study
For the early inflationary universe we can neglect the
the evolution of the matter, radiation and vacuum en-
constant term c0 , which is relevant only for the current
ergy densities in terms of the scale factor a. Applying
universe and essentially determines the value of the mea-
Bianchi’s identity ∇µ Gµν = 0 to Einstein’s equations
sured CC, see equation (4). Combining (6), (7) and sub-
with a CC term, Gµν = 8πG(Tµν + gµν ρΛ ), one finds [6]:
stituting (3) we find:
ρ̇ + 3H(ρ + p) = −ρ̇Λ , (5) 
H4

′ 2 2
aHH + 2H = 2 νH + α 2 . (14)
where we have used the FLRW metric to calculate the co- HI
variant derivative, and the EoS of vacuum ωΛ = pΛ /ρΛ =
Solving this equation leads [6]:
−1. Note that ρ = ρm + ρr and p refer respectively to
the density and pressure of both radiation and relativistic
r
1−ν HI
matter. We should also recall Friedmann’s equations: H(a) = √ , (15)
α 1 + Da4(1−ν)
3H 2 = 8πG(ρ + ρΛ ) , (6) where D must be a positive constant so that the Hubble
2
2Ḣ + 3H = −8πG(p + pΛ ) , (7) rate
p decreases with the expansion. Notice that H(0) =
(1 − ν)/αHI is finite (as long as α ̸= 0) and therefore
where we assume that the spatial curvature is zero or there is no singular initial point in the RVMs. That is,
negligible. if there were such an initial singularity, it was erased by
the process of inflation, together with the spatial curva-
ture and any other previous information of the Universe.
A. Current universe It is also important to note that in the initial period
Da4(1−ν) ≪ 1, H(a) ≃ H(0) and thus integrating the
As we have mentioned, the post-inflationary universe is scale factor we find a(t) = a(0)eH(0)t , i.e. the scale fac-
driven by the O(H 2 ) term. Rewriting (3) by substituting tor increases exponentially during the early inflationary
(4) we obtain: period.
Substituting (15) in Friedmann’s equations we find for
3ν
ρΛ (H) = ρ0Λ + (H 2 − H02 ) . (8) the vacuum and radiation density [5]:
8πG
(1 − ν)â4(1−ν)
Writing (5) for either a matter or radiation-dominated ρr (â) = ρ̂I , (16)
density, with ωm = 0 or ωr = 1/3: [1 + â4(1−ν) ]2
1 + νâ4(1−ν)
3 ρΛ (â) = ρ̂I , (17)
ρ + (1 + ω)ρ = −ρ′Λ ,
′
(9) [1 + â4(1−ν) ]2
a

Treball de Fi de Grau 2 Barcelona, June 2022

Vacuum energy and cosmological inflation Daniel Tuells Miralles

where we have re-scaled ĤI = H(0), ρ̂I = 3ĤI2 /8πG A. Current universe
and â ≡ a/a∗ . Here a∗ ≡ D−1/[4(1−ν)] essentially corre-
sponds to the vacuum-radiation equality point (not to be By simply substituting (13) in (19) we find:
mistaken with the radiation-matter equality point which
π
occurs in a much later stage of the cosmic evolution). We SA (a) = h i. (20)
Ω0m
can immediately see that ρr (0) = 0 and ρΛ (0) = ρ̂I , thus GH02 1 + 1−ν (a
−3(1−ν) − 1)
in the beginning there is only vacuum energy. Vacuum
then decays into radiation, however ρΛ remains essen- To determine SV we must revisit equation (10). Since
tially constant in the very first instants, causing a short ρm = mn, if we assume the particle mass remains con-
inflationary period. The radiation density continues to stant during the cosmic history m = m0 , then the
increase until the equality point a∗ is reached, after which number density cannot follow the standard dilution law
it starts to decay as well. It is also important to realise (n ∼ a−3 ) but rather an ”anomalous” one:
that |ρΛ /ρr | ∝ |ν| ≪ 1 after the equality point, and
therefore we would have a smooth transit to the radia- n(a) = n0 a−3(1−ν) . (21)
tion epoch of the ΛCDM picture, where BBN can occur This equation, given ν > 0, implies particle production
[6]. All these results are represented in Figure 1. through vacuum decay is still active even in our current
universe. In our present universe the entropy will be
dominated by the matter contribution, thus:
4π 3 4πσ n(a)
SV = nσVh = lh nσ = , (22)
3 3 H 3 (a)
with σ being the specific particle entropy. For the sake
of simplicity we can set σ = kB (keeping in mind that
Boltzmann’s constant is the unit of entropy), and in nat-
ural units kB = 1. Using (21) and (13) we find:
4πn0 a−3(1−ν)
SV (a) = h i3/2 . (23)
Ω0m
3H03 1 + 1−ν (a
−3(1−ν) − 1)

Notice that SV (a → ∞) = 0, i.e. the volume entropy

is actually decreasing with the expansion, and therefore
cannot fulfil the GSL by itself. Its numerical significance,
FIG. 1: Evolution of ρΛ (â) and ρr (â) in the early universe
(normalised to ρI ), represented in double-logarithmic scale. however, is negligible in front of SA , since:
We set ν = 10−3 . n0 ρ0m 1

H0

1 1
∼ ∼ ∼ ≪ , (24)
H03 mH03 GmH0 m GH02 GH02
where we used H02 ∼ Gρ0m , from (6), and in the last
III. GENERALISED SECOND LAW inequality we considered H0 ∼ 10−42 GeV which is much
smaller than any known particle mass. Hence it seems SA
For our thermodynamic discussion we shall adopt the should be able to continue increasing the total entropy
perspective of the GSL, where one takes the apparent and bring the universe to thermodynamic equilibrium.
horizon as the characteristic length and studies the evo- Differentiating equations (20) and (23) one indeed finds
′ ′′
lution of the entropy inside it, SV , with contributions (see [5] for details): Stotal (a → ∞) > 0 and Stotal (a →
from both matter and radiation; as well as the entropy ∞) < 0. Thus the GSL is fulfilled despite the decreasing
of the horizon area, SA [12]. The following conditions SV . In Figure 2 we can observe these results. See also
must be fulfilled: [13] for a comparison between the GSL predictions and
observations, with positive conclusions.
SV′ + SA
′
≥ 0, SV′′ + SA
′′
< 0. (18)

The first condition ensures the increase of entropy while B. Early universe
the second guarantees the system will reach thermody-
namic equilibrium. In the case of a FLRW universe with To study the entropy evolution in the early universe
no spatial curvature, the apparent horizon has the form we must rewrite (15) using the re-scaled parameters we
lh (t) = H −1 (t), and surface entropy [5]: introduced in the previous section:

A πl2 π ĤI
SA = = h = . (19) H(â) = √ . (25)
4G G GH 2 1 + â4(1−ν)

Treball de Fi de Grau 3 Barcelona, June 2022

Vacuum energy and cosmological inflation Daniel Tuells Miralles

FIG. 2: Evolution of SA (a) for ν = 10−3 and ν = 10−4 , FIG. 3: Evolution of SA (â) and SV (â), for ν = 10−3 and
normalised to the current value, from the present universe ν = 10−4 , normalised to the value at the equality point, in the
into the future. Recall that SV is numerically insignificant. early inflationary universe. There is an unrestricted growth
′′
of entropy due to Stotal (â) > 0.

Substituting in (20):
π(1 + â4(1−ν) ) and the entropy is dominated by the cosmological con-
SA (â) = . (26) stant. Therefore, the presence of c0 , which is negligible
GĤI2 in the early universe but becomes the dominant term in
The entropy inside the horizon, on the other hand, the current universe, is responsible for decelerating the
will be determined by the radiation contribution Sr = entropy production and setting the universe on the path
pr +ρr 3 4 ρr 3 towards thermodynamic equilibrium and the fulfillment
Tr Vh = 3 Tr Vh , where the radiation temperature Tr
is related to its density via: ρr = π 2 g∗ Tr4 /30, with g∗ be- of the GSL.
ing a factor counting the number of effectively massless
degrees of freedom [5]. Using (16) and (25) we find:
IV. HORIZON AND ENTROPY
1/4
4 π 2 g∗

SV (â) = ρ3/4 3
r (â)Vh =
3 30 Next we will discuss the potential of the RVMs to solve
!3 the horizon and entropy problems afflicting the ΛCDM
8π 3 T̂I model, namely the apparent inability of the latter to ex-
= g∗ (1 − ν)3/4 â3(1−ν) . (27)
135 ĤI plain neither how the universe evolved from a low entropy
value to a large one (giving rise to the arrow of time) [6],
In the last expression we have conveniently defined T̂I = nor the homogeneity of the universe, as it predicts a de-
 1/4
30ρ̂I
. Notice that SA ∼ â4(1−ν) and SV ∼ â3(1−ν) creasing horizon moving into the past. Let us first define
π 2 g∗
the concept of particle horizon, the visible region for a
and thus we have a huge entropy production caused by co-moving observer at a given time [5]:
inflation. In the next section we shall see this result might
provide a solution to the ΛCDM entropy problem. Both Z t
dt
Z a
da′
the surface and volume entropy are represented in Figure lp (a) = a =a . (29)
ti a(t) 0 a′2 H(a′ )
3.
Differentiating equations (26) and (27) it can be seen
′ ′′ In the ΛCDM model the integral tends to 0 for a → 0
that (see [5] for details): Stotal (â) > 0 and Stotal (â) > 0.
both in the radiation and matter epochs, with H ∼ a−2
That is, the GSL is violated in the early universe. We
and H ∼ a−3/2 respectively. This means that the ob-
shouldn’t be too alarmed by this result as the cosmic
servers become isolated in the past, or in other words that
evolution will eventually slow down the growth of entropy
the number of causally disconnected regions increases. In
and lead to equilibrium in the final de Sitter epoch, as
the context of the RVMs, however, we can substitute (25)
we have seen in the previous section. In fact, rewriting
in the previous equation to find:
(20):
Z a ′p
π(1 − ν) lp (a) 1 da
SA (a) = , (28) = lim 1 + â′4(1−ν) → ∞ . (30)
H02 [Ω0m a−3(1−ν) + Ω0Λ − ν] a ϵ→0 ĤI ϵ a′2

we see that as the ∼ a−3(1−ν) term disappears with the Therefore, the particle horizon increases faster than the
expansion, the denominator becomes H02 (Ω0Λ − ν) = c0 scale factor and becomes infinite (i.e. there is no horizon

Treball de Fi de Grau 4 Barcelona, June 2022

Vacuum energy and cosmological inflation Daniel Tuells Miralles

at all) in the primitive universe. In conclusion, the RVMs verse, however, is driven by an O(H 4 ) term, which
seem to have no horizon problem. causes a short period of rapid inflation, with an
In the case of the entropy problem, in the context of the exponential increase of the scale factor. During in-
ΛCDM model the present value enclosed in our horizon flation, the vacuum energy decays into relativistic
is [6]: particles, increasing the radiation density ρr , un-
til the vacuum-radiation equality point is reached,
2π 2 after which they both start decaying (Figure 1).
S0 = 3
g∗,0 Tr0 H0−3 ∼ 1088 , (31)
45 Nevertheless, during this post-inflationary decay
|ρΛ /ρr | ≪ 1, and therefore the main features of
with Tr0 being the current temperature of the CMB. This the ΛCDM radiation epoch are preserved, such as
result evolves in time as (T H −1 )3 , and for an adiabatic the Big Bang Nucleosynthesis.
evolution T ∼ a−1 , so the entropy evolves as S ∼ a3 and
S ∼ a3/2 for the radiation and matter epochs, respec- • We introduced the Generalised Second Law (GSL),
tively [6]. At the time of recombination (arec ∼ 10−3 ), which studies the surface entropy of the apparent
for example, the entropy would be Srec ∼ 1083 , and at horizon as well as the entropy contained inside it;
the primordial nucleosynthesis epoch (aBBN ∼ 10−9 ) and tried to determine whether or not the RVMs
it would be SBBN ∼ 1061 . This would imply that comply with it. We found that it is fulfilled in
our present horizon contains ∼ 105 regions which were the present universe, i.e. the current universe is
causally disconnected at recombination, and ∼ 1027 at on the path towards thermodynamic equilibrium
primordial nucleosynthesis. We have stumbled upon a (Figure 2). On the other hand, it was not satis-
re-phrased horizon problem. In the context of the RVMs, fied in the early universe as there is an unlimited
however, there is a huge entropy production driven by the entropy growth (Figure 3). We saw, however, that
O(H 4 ) term during the early inflationary period, which this apparent violation of the GSL was not worry-
is then transferred to the radiation epoch and preserved ing at all, as the presence of the constant term c0
throughout the cosmic evolution, accounting for the enor- in ρΛ (H), which was negligible in the initial period
mous value at present (31). In summary, the RVMs might of the cosmic evolution but became the dominant
also provide a possible solution to the entropy problem. term in the late universe, would eventually slow
′′
down the entropy growth until Stotal < 0, setting
off towards equilibrium.
V. CONCLUSIONS
• We determined that this enormous entropy produc-
• In this work we have studied some aspects of the tion during inflation, which is preserved throughout
Running Vacuum Models (RVMs), we have solved the entire cosmic history up until the present, may
its cosmological equations and studied some of the be able to provide a solution to the ΛCDM entropy
thermodynamic implications from the early uni- problem. The RVMs might also solve the horizon
verse until the present time. We have explored also problem, as they have no particle horizon at all
its asymptotic regime, in particular we have con- in the primitive universe and hence there are no
sidered if the RVMs universe tends eventually to causality issues.
a state of thermodynamic equilibrium. From the
confrontation of the RVMs predictions with the ob-
servational data in the present universe, it has been Acknowledgments
suggested that the vacuum energy is not necessar-
ily a strict constant, but may be mildly dynamical I want to thank my advisor Dr. Joan Solà, as well as
due to the ∼ νH 2 term [6, 8, 9]. The early uni- my family and friends.

[1] V. Sahni, A. Shafieloo and A. A. Starobinsky, Astrophys. 80(8) (2009) 083511.

J. 793 (2014) L40. [8] A. Gómez-Valent, J. Solà and S. Basilakos, JCAP 1501
[2] J. Solà, J. Phys. Conf. Ser. 453 (2013) 012015. (2015) 004.
[3] C. Moreno-Pulido and J. Solà, Renormalizing the vac- [9] J. Solà, A. Gómez-Valent and J. de Cruz Pérez, Phys.
uum energy in cosmological spacetime: implications for Dark Univ. 25 (2019) 100311.
the cosmological constant problem (to appear in Euro- [10] S. Basilakos, M. Plionis and J. Solà, Phys.Rev. D82
phys. J. C, 2022). (2010) 083512.
[4] J. Solà, Phil. Trans. R. Soc. A 380 (2022) 20210182. [11] J. Solà, J. of Phys. A41 (2008) 164066.
[5] J. Solà and H. Yu, Gen. Rel. Grav. 52 (2020) 2,17. [12] G. W. Gibbons and S. W. Hawking, Phys. Rev. D 15
[6] J. Solà and A. Gómez-Valent, Int. J. Mod. Phys. D 24 (1977) 2738.
(2015) 1541003. [13] M. González-Espinoza and D. Pavón, MNRAS 484
[7] S. Basilakos, M. Plionis and J. Solà, Physical Review D (2019) 2924.

Treball de Fi de Grau 5 Barcelona, June 2022