Anda di halaman 1dari 15

Corfu Summer Institute on Elementary Particle Physics, 1998

PROCEEDINGS

Introduction to Cosmology

G. Lazarides
Physics Division, School of Technology,
Aristotle University of Thessaloniki,
Thessaloniki 540 06, Greece
E-mail: lazaride@eng.auth.gr
arXiv:hep-ph/9904502 v1 30 Apr 1999

Abstract: The standard big bang cosmological model and the history of the early universe accord-
ing to the grand unified theories of strong, weak and electromagnetic interactions are summarized.
The shortcomings of big bang are discussed together with their resolution by inflationary cosmol-
ogy. Inflation and the subsequent oscillation and decay of the inflaton field are studied. The density
perturbations produced during inflation and their evolution during the matter dominated era are an-
alyzed. The temperature fluctuations of the cosmic background radiation are summarized. Finally,
the nonsupersymmetric as well as the supersymmetric hybrid inflationary model is described.

1. The Big Bang Model based, is that the universe is homogeneous and
isotropic. The strongest evidence so far for this
The discovery of the cosmic background radia- cosmological principle is the observed [2] isotropy
tion (CBR) in 1964 together with the observed of the CBR. Under this assumption, the four di-
Hubble expansion of the universe had established mensional spacetime in the universe is described
hot big bang cosmology as a viable model of by the Robertson-Walker metric
the universe. The success of the theory of nu-
cleosynthesis in reproducing the observed abun- ds2 = −dt2 +
dance pattern of light elements together with the
 
proof of the black body character of the CBR 2 dr2
a (t) + r2(dθ2 + sin2 θ dϕ2) , (1.1)
then established hot big bang as the standard 1 − kr2
cosmological model. This model combined with
where r, ϕ and θ are ‘comoving’ polar coordi-
grand unified theories (GUTs) of strong, weak
nates, which remain fixed for objects that have
and electromagnetic interactions provides an ap-
no other motion than the general expansion of
propriate framework for discussing the very early
the universe. The parameter k is the ‘scalar cur-
stages of the universe evolution. A brief intro-
vature’ of the 3-space and k = 0, k > 0 or k < 0
duction to hot big bang follows.
correspond to flat, closed or open universe. The
1.1 Hubble Expansion dimensionless parameter a(t) is the ‘scale factor’
of the universe and describes cosmological expan-
For cosmic times t ∼ >
tP ≡ MP−1 ∼ 10−44 sec
19 sion. We normalize it by taking a0 ≡ a(t0 ) = 0,
(MP = 1.22 × 10 GeV is the Planck scale) af-
where t0 is the present cosmic time.
ter the big bang, quantum fluctuations of grav-
The ‘instantaneous’ radial physical distance
ity cease to exist. Gravitation can then be ade-
is given by
quately described by classical relativity. Strong,
weak and electromagnetic interactions, however, Z r
dr
require relativistic quantum field theoretic treat- R = a(t) · (1.2)
0 (1 − kr2 )1/2
ment and are described by gauge theories.
An important principle, on which the stan- For flat universe (k = 0), R̄ = a(t)r̄ (r̄ is a ‘co-
dard big bang (SBB) cosmological model [1] is moving’ and R̄ a physical vector in 3-space) and
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

the velocity of an object is Averaging p, we write ρ + p = γρ. Eq.(1.5)


then becomes ρ̇ = −3Hγρ, which gives dρ/ρ =
dR̄ ȧ dr̄
V̄ = = R̄ + a , (1.3) −3γda/a and ρ ∝ a−3γ . For a universe domi-
dt a dt
nated by pressureless matter, p = 0 and, thus,
where overdots denote derivation with respect to γ = 1, which gives ρ ∝ a−3 . This is easily in-
cosmic time. The first term in the right hand side terpreted as mere dilution of a fixed number of
(rhs) of this equation is the ‘peculiar velocity’, particles in a ‘comoving’ volume due to the cos-
v̄ = a(t)r̄˙ , of the object, i.e., its velocity with mological expansion. For a radiation dominated
respect to the ‘comoving’ coordinate system. For universe, p = 1/3 and, thus, γ = 4/3, which gives
v̄ = 0, Eq.(1.3) becomes ρ ∝ a−4. In this case, we get an extra factor of
ȧ a(t) due to the red-shifting of all wave-lengths by
V̄ = R̄ ≡ H(t)R̄ , (1.4)
a the expansion. Substituting ρ ∝ a−3γ in Fried-
mann equation with k = 0, we get ȧ/a ∝ a−3γ/2
where H(t) ≡ ȧ(t)/a(t) is the Hubble parame-
and, thus, a(t) ∝ t2/3γ . Taking into account the
ter. This is the well-known Hubble law asserting
normalization of a(t) (a(t0 ) = 1), this gives
that all objects run away from each other with
velocities proportional to their distances and is a(t) = (t/t0)2/3γ . (1.9)
considered as the first success of SBB cosmology.
For a matter dominated universe, we get the ex-
1.2 Friedmann Equation pansion law a(t) = (t/t0)2/3 . ‘Radiation’, how-
ever, expands as a(t) = (t/t0 )1/2.
Homogeneity and isotropy of the universe imply
The universe in its early stages of evolution
that the energy momentum tensor takes the diag-
is radiation dominated and its energy density is
onal form (Tµν ) = diag(−ρ, p, p, p), where ρ is the
 
energy density of the universe and p the pressure. π2 7
ρ= Nb + Nf T 4 ≡ c T 4 , (1.10)
Energy momentum conservation (Tµν;ν = 0) then 30 8
takes the form of the continuity equation where T is the cosmic temperature and Nb (Nf )
dρ is the number of massless bosonic (fermionic) de-
= −3H(t)(ρ + p) , (1.5)
dt grees of freedom. The combination g∗ = Nb +
where the first term in the rhs describes the di- (7/8)Nf is called effective number of massless de-
lution of the energy due to the expansion of the grees of freedom. The entropy density is
universe and the second term corresponds to the 2π2
s= g∗ T 3 . (1.11)
work done by pressure. Eq.(1.5) can be given the 45
following more transparent form Assuming adiabatic universe evolution, i.e., con-
 
4π 3 stant entropy in a ‘comoving’ volume (sa3 =
d a ρ = −p 4πa2da , (1.6) constant), we obtain the relation aT = constant.
3
The temperature-time relation during radiation
which indicates that the energy loss of a ‘comov-
dominance is then derived from Friedmann equa-
ing’ sphere of radius ∝ a(t) equals the work done
tion (with k = 0):
by pressure on its boundary as it expands.
For a universe described by the Robertson- MP
T2 = · (1.12)
Walker metric in Eq.(1.1), Einstein’s equations 2(8πc/3)1/2t

1 ν We see that classically the expansion starts at


Rµν − δ R = 8πG Tµν , (1.7) t = 0 with T = ∞ and a = 0. This initial singu-
2 µ
larity is, however, not physical since general rel-
where Rµν and R are the Ricci tensor and scalar
ativity fails at cosmic times smaller than about
curvature tensor and G ≡ MP−2 is the Newton’s
the Planck time tP . The only meaningful state-
constant, lead to the Friedmann equation
ment is that the universe, after a yet unknown
 2
ȧ(t) 8πG k initial stage, emerges at a cosmic time ∼ tP with
H2 ≡ = ρ− 2 · (1.8) temperature T ∼ MP .
a(t) 3 a

2
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

1.3 Important Cosmological Parameters For ‘matter’ (‘radiation’), this becomes dH (t) =
The most important parameters describing the 3t = 2H −1(t) (dH (t) = 2t = H −1(t)). The
expanding universe are the following: present particle horizon is dH (t0 ) = 2H0−1 ≈
6, 000 h−1 Mpc, the present cosmic time is t0 =
i. The present value of the Hubble parameter 2H0−1/3 ≈ 6.7 × 109 h−1 years and the present
(known as Hubble constant) H0 ≡ H(t0 ) = value of the critical density is ρc = 3H02/8πG ≈
100 h km sec −1 Mpc−1 (0.4 ∼<
h∼<
0.8). 1.9 × 10−29 h−2 gm/cm3.
ii. The fraction Ω = ρ/ρc , where ρc is the
critical density corresponding to a flat uni- 1.5 Brief History of the Early Universe
verse (k = 0). From Friedmann equation,
We will now briefly describe the early stages of
ρc = 3H 2/8πG and, thus, Ω = 1+k/a2 H 2.
the universe evolution according to GUTs [4].
Ω = 1, Ω > 1 or Ω < 1 correspond to
We will take a GUT based on the gauge group G
flat, closed or open universe. Assuming in-
(= SU (5), SO(10), SU (3)3 , ...) with or without
flation (see below), the present value of Ω
supersymmetry. At a superheavy scale MX ∼
must be Ω0 = 1. However, the baryonic
1016 GeV (the GUT mass scale), G breaks to
contribution to Ω is ΩB ≈ 0.05 − 0.1 [3].
the standard model gauge group GS = SU (3)c ×
This indicates that most of the energy in
SU (2)L ×U (1)Y by the vacuum expectation value
the universe must be in nonbaryonic form.
(vev) of an appropriate higgs field φ. (For sim-
iii. The deceleration parameter plicity, we will consider that this breaking occurs
(ä/ȧ) ρ + 3p in just one step.) GS is, subsequently, broken to
q=− = · (1.13) SU (3)c × U (1)em at the electroweak scale MW .
(ȧ/a) 2ρc
GUTs together with the SBB cosmological
For ‘matter’, q = Ω/2 and, thus, inflation model (based on classical gravitation) provide a
implies that the present deceleration pa- suitable framework for discussing the early his-
rameter is q0 = 1/2. tory of the universe for cosmic times ∼>
10−44 sec.
1.4 Particle Horizon They predict that the universe, as it expands and
cools down after the big bang, undergoes [5] a se-
Light travels only a finite distance from the time ries of phase transitions during which the initial
of big bang (t = 0) till some cosmic time t. From gauge symmetry is gradually reduced and several
the Robertson-Walker metric in Eq.(1.1), we find
important phenomena take place.
that the propagation of light along the radial di-
After the big bang, the GUT gauge group
rection is described by the equation a(t)dr = dt.
G was unbroken and the universe was filled with
The particle horizon, which is the ‘instantaneous’
a hot ‘soup’ of massless particles which included
distance at time t travelled by light since the be-
not only photons, quarks, leptons and gluons but
ginning of time, is then given by
also the weak gauge boson W ± , Z 0 , the GUT
Z t
dt0 gauge bosons X, Y , ... as well as several higgs
dH (t) = a(t) 0
· (1.14)
0 a(t ) bosons. (In the supersymmetric case, all the
The particle horizon is a very important notion supersymmetric partners of these particles were
since it coincides with the size of the universe also present.) At cosmic time t ∼ 10−37 sec
already seen at time t or, equivalently, with the corresponding to temperature T ∼ 1016 GeV,
distance at which causal contact has been estab- G broke down to GS and the X, Y , ... gauge
lished at t. Eqs.(1.9) and (1.14) give bosons together with some higgs bosons acquired
superheavy masses of order MX . The out-of-

dH (t) = t , γ 6= 2/3 . (1.15) equilibrium decay of these superheavy particles
3γ − 2 can produce [6] the observed baryon asymme-
Also, try of the universe (BAU). Important ingredients
2 −1 2 for this mechanism to work are the violation of
H(t) = t , dH (t) = H −1(t) . (1.16) baryon number, which is inherent in GUTs, and
3γ 3γ − 2

3
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

C and CP violation. This is the second impor- bang. Finally, structure formation [15] in the
tant success of the SBB model. universe starts at t ≈ 2 × 108 years.
During the GUT phase transition, topologi-
cally stable extended objects [7] such as magnetic
monopoles [8], cosmic strings [9] or domain walls
2. Shortcomings of Big Bang
[10] can also be produced. Monopoles, which ex-
The SBB cosmological model has been very suc-
ist in all GUTs, can lead into cosmological prob-
cessful in explaining, among other things, the
lems [11] which are, however, avoided by inflation
Hubble expansion of the universe, the existence
[12, 13] (see Secs.2.3 and 3.3). This is a period of
of the CBR and the abundances of the light ele-
an exponentially fast expansion of the universe
ments which were formed during primordial nu-
which can occur during some GUT phase transi-
cleosynthesis. Despite its great successes, this
tion. Strings can contribute [14] to the primor-
model had a number of long-standing shortcom-
dial density fluctuations necessary for structure
ings which we will now summarize:
formation [15] in the universe whereas domain
walls are [10] absolutely catastrophic and GUTs
2.1 Horizon Problem
predicting them should be avoided or inflation
should be used to remove them from the scene. The CBR, which we receive now, was emitted at
At t ∼ 10−10 sec or T ∼ 100 GeV, the elec- the time of ‘decoupling’ of matter and radiation
troweak transition takes place and GS breaks to when the cosmic temperature was Td ≈ 3, 000 K.
SU (3)c × U (1)em . The W ± , Z 0 gauge bosons to- The decoupling time, td , can be calculated from
gether with the electroweak higgs fields acquire  2/3
masses ∼ MW . Subsequently, at t ∼ 10−4 sec T0 2.73 K a(td ) td
= = = · (2.1)
or T ∼ 1 GeV, color confinement sets in and the Td 3, 000 K a(t0 ) t0
quarks get bounded forming hadrons.
It turns out that td ≈ 200, 000 h−1 years.
The direct involvement of particle physics es-
The distance over which the photons of the
sentially ends here since most of the subsequent
CBR have travelled since their emission is
phenomena fall into the realm of other branches.
Z t0 "  2/3#
We will, however, sketch some of them since they dt0 td
are crucial for understanding the earlier stages of a(t0 ) 0
= 3t0 1 −
td a(t ) t0
the universe evolution where their origin lies.
At t ≈ 180 sec (T ≈ 1 MeV), nucleosynthesis
≈ 3t0 ≈ 6, 000 h−1 Mpc , (2.2)
takes place, i.e., protons and neutrons form nu-
clei. The abundance of light elements (D, 3 He, which essentially coincides with the present par-
4
He and 7 Li) depends [16] crucially on the num- ticle horizon size. A sphere around us with ra-
ber of light particles (with mass ∼ <
1 MeV), i.e., dius equal to this distance is called the ‘last scat-
the number of light neutrinos, Nν , and ΩB h2. tering surface’ since the CBR observed now has
Agreement with observations [3] is achieved for been emitted from it. The particle horizon size
Nν = 3 and ΩB h2 ≈ 0.019. This is the third at td was 2H −1(td ) = 3td ≈ 0.168 h−1 Mpc and
success of SBB cosmology. Much later, at the expanded till the present time to become equal
so called ‘equidensity’ point, teq ≈ 3, 000 years, to 0.168 h−1 (a(t0 )/a(td )) Mpc ≈ 184 h−1 Mpc.
‘matter’ dominates over ‘radiation’. The angle subtended by this ‘decoupling’ horizon
At cosmic time t ≈ 200, 000 h−1 years (T ≈ at present is θd ≈ 184/6, 000 ≈ 0.03 rads ≈ 2 o.
3, 000 K), we have the ‘decoupling’ of ‘matter’ Thus, the sky splits into 4π/(0.03)2 ≈ 14, 000
and ‘radiation’ and the ‘recombination’ of atoms. patches that never communicated causally before
After this, ‘radiation’ evolves as an independent sending light to us. The question then arises how
(not interacting) component of the universe and come the temperature of the black body radia-
is detected today as CBR with temperature T0 ≈ tion from all these patches is so accurately tuned
2.73 K. The existence of this radiation is the as the measurements of the cosmic background
fourth important success of the theory of big explorer [2] (COBE) require (δT /T ≈ 6.6×10−6).

4
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

2.2 Flatness Problem the higgs correlation length ξ(T ) = m−1 H (T ) ex-
ceeds T . This condition reads (4π/3)ξ 3∆V ∼ >
T,
The present energy density, ρ, of the universe 2 4
where ∆V ∼ λ hφi is the difference in free en-
has been observed to lie in the relatively nar-
ergy density between φ = 0 and φ = hφi(T ). The
row range 0.1ρc ∼ <
ρ∼
<
2ρc , where ρc is the crit-
Ginzburg temperature [17], TG , corresponds to
ical energy density corresponding to a flat uni-
the saturation of this inequality. So, at T ∼<
TG ,
verse. The lower bound has been derived from
the fluctuations over φ = 0 stop and hφi settles
estimates of galactic masses using the virial the-
on the vacuum manifold G/GS . At TG , the uni-
orem whereas the upper bound from the volume
verse splits into regions of size ξG ∼ (λ2Tc )−1,
expansion rate implied by the behavior of galac-
the higgs correlation length at TG , with the higgs
tic number density at large distances. Eq.(1.8)
field being more or less aligned in each region.
implies that (ρ − ρc )/ρc = 3(8πGρc)−1 (k/a2 )
Monopoles are produced at the corners where
is proportional to a, for matter dominated uni-
such regions meet (Kibble [7] mechanism) and
verse. Consequently, in the early universe, we
their number density is estimated to be nM ∼
have |(ρ − ρc )/ρc |  1 and the question arises −3
pξG ∼ pλ6 Tc3, where p ∼ 1/10 is a geometric
why the initial energy density of the universe was
factor. The ‘relative’ monopole number density
so finely tuned to be equal to its critical value.
then turns out to be rM = nM /T 3 ∼ 10−6. We
can derive a lower bound on rM by employing
2.3 Magnetic Monopole Problem
causality. The higgs field φ cannot be correlated
This problem arises only if we combine the SBB at distances bigger than the particle horizon size,
model with GUTs [4] of strong, weak and elec- 2tG, at TG . This gives the causality bound
tromagnetic interactions. As already indicated, p
according to GUTs, the universe underwent [5] nM >
∼ 4π , (2.3)
3
(2tG )3
a phase transition during which the GUT gauge
symmetry group, G, broke to GS . This breaking which implies that rM ∼ >
10−10.
was due to the fact that, at a critical tempera- The subsequent evolution of monopoles, af-
ture Tc , an appropriate higgs field, φ, developed ter TG , is governed by the equation [11]
a nonzero vev. Assuming that this phase transi-
tion was a second order one, we have hφi(T ) ≈ dnM ȧ
= −Dn2M − 3 nM , (2.4)
hφi(T = 0)(1 − T 2 /Tc2)1/2, mH (T ) ≈ λhφi(T ), dt a
for the temperature dependent vev and mass of where the first term in the rhs (with D being
the higgs field respectively at T ≤ Tc (λ is an an appropriate constant) describes the dilution
appropriate higgs coupling constant). of monopoles due to their annihilation with an-
The GUT phase transition produces mag- timonopoles while the second term corresponds
netic monopoles [8] which are localized devia- to their dilution by the general cosmological ex-
−1
tions from the vacuum with radius ∼ MX , en- pansion. The monopole-antimonopole annihila-
ergy ∼ MX /αG and φ = 0 at their center (αG = tion proceeds as follows. Monopoles diffuse to-
2
gG /4π with gG being the GUT gauge coupling wards antimonopoles in the plasma of charged
constant). The vev of the higgs field on a sphere, particles, capture each other in Bohr orbits and
−1
S 2 , with radius  MX around the monopole eventually annihilate. The annihilation is effec-
lies on the vacuum manifold G/GS and we, thus, tive provided the mean free path of monopoles
obtain a mapping: S 2 −→ G/GS . If this map- in the plasma of charged particles does not ex-
ping is homotopically nontrivial the topological ceed their capture distance. This happens at
stability of the monopole is guaranteed. cosmic temperatures T ∼ >
1012 GeV. The over-
Monopoles can be produced when the fluc- all result is that, if the initial relative magnetic
tuations of φ over φ = 0 between the vacua at monopole density rM,in ∼ >
10−9(∼ <
10−9), the fi-
−9
±hφi(T ) cease to be frequent. This takes place nal one rM,fin ∼ 10 (∼ rM,in ). This combined
when the free energy needed for φ to fluctuate with the causality bound yields rM,fin ∼ >
10−10.
from hφi(T ) to zero in a region of radius equal to However, the requirement that monopoles do not

5
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

dominate the energy density of the universe at Now let us take H 6= 0. Since the cosmolog-
nucleosynthesis gives ical expansion pulls the particles apart, we get a
smaller growth:
rM (T ≈ 1 MeV) ∼
<
10−19 , (2.5)
δk̄ ∝ a(t) ∝ t2/3 , (2.9)
and we obtain a clear discrepancy of about ten
orders of magnitude. in the matter dominated case. For a radiation
dominated universe (p 6= 0), we get essentially no
2.4 Density Fluctuations growth of the density perturbations. This means
that, in order to have structure formation in the
For structure formation [15] in the universe, we
universe, which requires δρ/ρ ∼ 1, we must have
need a primordial density perturbation, δρ/ρ, at
all length scales with a nearly flat spectrum [18]. δρ
( )eq ∼ 4 × 10−5(Ω0 h)−2 , (2.10)
We also need some explanation of the temper- ρ
ature fluctuations, δT /T , of CBR observed by at the ‘equidensity’ point (where the energy den-
COBE [2] at angles θ ∼ >
θd ≈ 2 o which violate sities of matter and radiation coincide), since the
causality (see Sec.2.1). available growth factor for perturbations is given
Let us expand δρ/ρ in plane waves by a0/aeq ∼ 2.5 × 104(Ω0 h)2. Here Ω0 = ρ0 /ρc ,
Z where ρ0 is the present energy density of the uni-
δρ
(r̄, t) = d3kδk̄ (t)eik̄r̄ , (2.6) verse. The question then is where these primor-
ρ
dial density fluctuations originate from.
where r̄ is a ‘comoving’ vector in 3-space and k̄
is the ‘comoving’ wave vector with k = |k̄| be- 3. Inflation
ing the ‘comoving’ wave number (λ = 2π/k is
the ‘comoving’ wave length whereas the physical Inflation [12, 13] is an idea which solves simulta-
wave length is λphys = a(t)λ). For λphys ≤ H −1, neously all four cosmological puzzles and can be
the time evolution of δk̄ is described by the New- summarized as follows. Suppose there is a real
tonian equation scalar field φ (the inflaton) with (symmetric) po-
tential energy density V (φ) which is quite ‘flat’
vs2 k2 near φ = 0 and has minima at φ = ±hφi with
δ̈k̄ + 2H δ̇k̄ + δ = 4πGρδk̄ , (2.7)
a2 k̄ V (±hφi) = 0. At high enough T ’s, φ = 0 in the
where the second term in the left hand side (lhs) universe due to the temperature corrections in
comes from the cosmological expansion and the V (φ). As T drops, the effective potential den-
third is the ‘pressure’ term (vs is the velocity of sity approaches the T =0 potential but a little
sound given by vs2 = dp/dρ, where p is the mean potential barrier separating the local minimum
pressure). The rhs of this equation corresponds at φ = 0 and the vacua at φ = ±hφi still re-
to the gravitational attraction. mains. At some point, φ tunnels out to φ1  hφi
For the moment, let us put H=0 (static uni- and a bubble with φ = φ1 is created in the uni-
verse). In this case, there exists a characteris- verse. The field then rolls over to the minimum
tic wave number kJ , the Jeans wave number, of V (φ) very slowly (due to the flatness of the
given by kJ2 = 4πGa2ρ/vs2 and having the fol- potential). During this slow roll over, the energy
lowing property. For k ≥ kJ , pressure dominates density ρ ≈ V (φ = 0) ≡ V0 remains essentially
over gravitational attraction and the density per- constant for quite some time. The Lagrangian
turbations just oscillate, whereas, for k ≤ kJ , density
1
gravitational attraction dominates and the den- L = ∂µ φ∂ µ φ − V (φ) (3.1)
2
sity perturbations grow exponentially. In partic-
gives the energy momentum tensor
ular, for p=0 (matter domination), vs = 0 and  
all scales are Jeans unstable with 1
Tµν = −∂µ φ∂ ν φ + δµν ∂λ φ∂ λ φ − V (φ) ,
2
δk̄ ∝ exp(t/τ ) , τ = (4πGρ)−1/2 . (2.8) (3.2)

6
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

which during the slow roll over takes the form Tr ∼ 109 GeV [22]. The universe, after that,
Tµν ≈ −V0 δµν . This means that ρ ≈ −p ≈ goes back to normal big bang cosmology. The
V0 , i.e., the pressure p is negative and equal in horizon d(tf ) is stretched during the period of φ-
magnitude with the energy density ρ, which is oscillations by some factor ∼ 109 depending on
consistent with Eq.(1.5). Since, as we will see, details and between Tr and the present era by
a(t) grows very fast, the ‘curvature’ term, k/a2, a factor Tr /T0. So it finally becomes equal to
in Eq.(1.8) becomes subdominant and we get H −1eHτ 109(Tr /T0), which should exceed 2H0−1
 2 in order to solve the horizon problem. Taking
2 ȧ 8πG V0 ≈ M X 4
, MX ∼ 1016 GeV, we see that, with
H ≡ = V0 , (3.3)
a 3 N = Hτ ∼ >
55, the horizon problem is evaded.
which gives a(t) ∝ eHt , H 2 = (8πG/3)V0 = con- 3.2 Resolution of the Flatness Problem
stant. So the bubble expands exponentially for
some time and a(t) grows by a factor The ‘curvature’ term of the Friedmann equation,
at present, is given by
a(tf )    −13 2
= expH(tf − ti ) ≡ expHτ , (3.4) k k
a(ti) −2Hτ −18 10 GeV
≈ e 10 ,
a2 a2 bi 109 GeV
between an initial (ti ) and a final (tf ) time. (3.6)
The inflationary scenario just described here, where the terms in the rhs correspond to the ‘cur-
known as new [19] inflation (with the inflaton vature’ term before inflation, and its growth fac-
field starting from the origin, φ=0), is certainly tors during inflation, during φ-oscillations and af-
not the only realization of the idea of inflation. ter ‘reheating’ respectively. Assuming (k/a2 )bi ∼
Another interesting possibility is to consider the (8πG/3)ρ ∼ H 2 (ρ ≈ V0), we get k/a20H02 ∼
universe as it emerges at the Planck time tP , 1048 e−2Hτ which gives (ρ0 − ρc )/ρc ≡ Ω0 − 1 =
where the fluctuations of gravity cease to exist. k/a20H02  1, for Hτ  55. In fact, strong infla-
We can imagine a region of size `P ∼ MP−1 where tion implies that the present universe is flat with
the inflaton field acquires a large and almost uni- a great accuracy.
form value and carries negligible kinetic energy.
Under certain circumstances this region can in- 3.3 Resolution of the Monopole Problem
flate (exponentially expand) as φ rolls down to- It is obvious that, with a number of e-foldings
wards its vacuum value. This type of inflation >
55, the primordial monopole density is diluted

with the inflaton starting from large values is by at least 70 orders of magnitude and they be-
known as the chaotic [20] inflationary scenario. come totally irrelevant. Also, since Tr  mM
We will now show that, with an adequate (=the monopole mass), there is no production of
number of e-foldings, N = Hτ , the first three magnetic monopoles after ‘reheating’.
cosmological puzzles are easily resolved (we leave
the question of density perturbations for later).
4. Detailed Analysis of Inflation
3.1 Resolution of the Horizon Problem
The Hubble parameter is not exactly constant
The particle horizon during inflation (exponen- during inflation as we, naively, assumed so far.
tial expansion) It actually depends on the value of φ:
Z t
dt0 8πG
d(t) = eHt Ht0
≈ H −1expH(t − ti ) , (3.5) H 2 (φ) = V (φ) . (4.1)
ti e 3
To find the evolution equation for φ during infla-
for t − ti  H −1, grows as fast as a(t). At the
tion, we vary the action
end of inflation (t = tf ), d(tf ) ≈ H −1 expHτ
Z p  
and the field φ starts oscillating about the min- 1
imun of the potential at φ = hφi. It then decays −det(g) d4x ∂µ φ∂ µ φ − V (φ) + M (φ) ,
2
and ‘reheats’ [21] the universe at a temperature (4.2)

7
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

Z φf
where g is the metric tensor and M (φ) represents V (φ)dφ 4πG 2
−8πG = (φi − φ2f ) . (4.9)
the coupling of φ to ‘light’ matter causing its φi V 0 (φ) ν
decay. We find Assuming that φi  φf , this reduces to N (φ) =
0 (4πG/ν)φ2.
φ̈ + 3H φ̇ + Γφ φ̇ + V (φ) = 0 , (4.3)

where the prime denotes derivation with respect 5. Coherent Field Oscillations
to φ and Γφ is the decay width [23] of the inflaton.
Assume, for the moment, that the decay time After the end of inflation at cosmic time tf , the
of φ, td = Γ−1 −1 term φ̈ takes over and Eq.(4.4) reduces to φ̈ +
φ , is much greater than H , the
expansion time for inflation. Then the term Γφ φ̇ V 0 (φ) = 0, which means that φ starts oscillating
can be ignored and Eq.(4.3) reduces to coherently about the global minimum of the po-
tential. In reality, due to the ‘friction’ term, φ
φ̈ + 3H φ̇ + V 0 (φ) = 0 . (4.4) performs damped oscillations with a rate of en-
ergy density loss given by
Inflation is by definition the situation where φ̈ is  
subdominant to the ‘friction’ term 3H φ̇ in this d 1 2
ρ̇ = φ̇ + V (φ) = −3H φ̇2 = −3H(ρ+p) ,
equation (and the kinetic energy density is sub- dt 2
dominant to the potential one). Eq.(4.4) then (5.1)
further reduces to the inflationary equation [24] where ρ = φ̇2/2 + V (φ) and the pressure p =
φ̇2/2 − V (φ). Averaging p over one oscillation of
3H φ̇ = −V 0(φ) , (4.5) φ and writing [25] ρ + p = γρ, we get ρ ∝ a−3γ
which gives and a(t) ∝ t2/3γ (see Sec.1.2).
The number γ for an oscillating field can be
V 00(φ)φ̇ V 0(φ) 0 written as (assuming a symmetric potential)
φ̈ = − + H (φ)φ̇ . (4.6)
3H(φ) 3H 2(φ) RT 2 R φmax
0
φ̇ dt 0
φ̇dφ
Comparing the two terms in the rhs of this equa- γ = RT = R φmax , (5.2)
ρdt (ρ/φ̇)dφ
tion with the ‘friction’ term in Eq.(4.4), we get 0 0

the conditions for inflation (slow roll conditions): where T and φmax are the period and the am-
 2 plitude of the oscillation respectively. From the
M 2 V 00(φ) MP2 V 0(φ) equation ρ = φ̇2/2 + V (φ) = Vmax , where Vmax
η ≡ P ≤ 1 ,  ≡ ≤1.
8π V (φ) 16π V (φ) is the maximal potential energy density, we ob-
(4.7) p
tain φ̇ = 2(Vmax − V (φ)). Substituting this in
The end of the slow roll over occurs when either
Eq.(5.2) we get [25]
of the these inequalities is saturated. If φf is Rφ
the value of φ at the end of inflation, then tf ∼ 2 0 max (1 − V /Vmax )1/2dφ
γ = R φmax · (5.3)
H −1 (φf ). (1 − V /Vmax )−1/2dφ
0
The number of e-foldings during inflation can
For a potential of the simple form V (φ) = λφν , γ
be calculated as follows:
  Z tf is readily found to be given by γ = 2ν/(ν + 2).
a(tf ) Consequently, in this case, ρ ∝ a−6ν/(ν+2) and
N (φi → φf ) ≡ `n = Hdt =
a(ti ) ti a(t) ∝ t(ν+2)/3ν . For ν = 2, in particular, one
Z φf Z φf has γ=1, ρ ∝ a−3 , a(t) ∝ t2/3 and the oscillating
H(φ) 3H 2(φ)dφ
dφ = − , (4.8) field behaves like pressureless ‘matter’. This is
φi φ̇ φi V 0(φ)
not unexpected since a coherent oscillating mas-
where Eqs.(3.4), (4.5) and the definition of H =
sive free field corresponds to a distribution of
ȧ/a were used. For simplicity, we can shift the
static massive particles. For ν=4, however, we
field φ so that the global minimum of the poten-
obtain γ = 4/3, ρ ∝ a−4 , a(t) ∝ t1/2 and the
tial is displaced at φ = 0. Then, if V (φ) = λφν system resembles ‘radiation’. For ν = 6, one has
during inflation, we have
γ = 3/2, ρ ∝ a−4.5, a(t) ∝ t4/9 and the expansion
Z φf
3H 2(φ)dφ is slower than in a radiation dominated universe
N (φi → φf ) = − =
φi V 0(φ) (the pressure is higher than in ‘radiation’).

8
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

6. Decay of the Field φ with ρ = ρf (t/tf )−2 exp(−Γφ t) being the energy
density of the field φ which performs damped os-
Reintroducing the ‘decay’ term Γφ φ̇, Eq.(4.3) can cillations and decays into ‘light’ particles.
be written as The energy density of the ‘new radiation’
 
d 1 2 grows relative to the energy density of the oscil-
ρ̇ = φ̇ + V (φ) = −(3H + Γφ )φ̇2 , (6.1)
dt 2 lating field and becomes essentially equal to it at
a cosmic time td = Γ−1 φ as one can deduce from
which is solved [21, 25] by
Eq.(6.7). After this time, the universe enters into
 −3γ
a(t) the radiation dominated era and the normal big
ρ(t) = ρf exp[−γΓφ (t − tf )] , (6.2) bang cosmology is recovered. The temperature
a(tf )
at td , Tr (td ), is historically called the ‘reheat’
where ρf is the energy density at the end of in- temperature although no supercooling and sub-
flation at cosmic time tf . The second and third
sequent reheating of the universe actually takes
factors in the rhs of this equation represent the
place. Using the time to temperature relation in
dilution of the field energy due to the expansion
Eq.(1.12) for a radiation dominated universe we
of the universe and the decay of φ to light parti-
find that
cles respectively.
 1/4
All pre-existing ‘radiation’ (known as ‘old ra- 45
Tr = (Γφ MP )1/2 , (6.8)
diation’) was diluted by inflation, so the only ‘ra- 16π3g∗
diation’ present is the one produced by the decay
where g∗ is the effective number of degrees of
of φ and is known as ‘new radiation’. Its energy
freedom. For a potential of the type V (φ) = λφν ,
density satisfies [21, 25] the equation
the total expansion of the universe during the
ρ̇r = −4Hρr + γΓφ ρ , (6.3) period of damped field oscillations is
  ν+2
where the first term in the rhs represents the di- a(td ) td 3ν

lution of radiation due to the cosmological expan- = . (6.9)


a(tf ) tf
sion while the second one is the energy density
transfer from φ to ‘radiation’. Taking ρr (tf )=0, 7. Density Perturbations
this equation gives [21, 25]
 −4 We are ready to sketch how inflation solves the
a(t) density fluctuation problem described in Sec.2.4.
ρr (t) = ρf
a(tf ) As a matter of fact, inflation not only homoge-
Z t  0 4−3γ nizes the universe but also provides us with the
a(t ) 0
e−γΓφ (t −tf ) γΓφ dt0 . (6.4) primordial density fluctuations necessary for the
tf a(t f )
structure formation in the universe. To under-
For tf  td and ν = 2, this expression is approx- stand the origin of these fluctuations, we must
imated by first introduce the notion of ‘event horizon’. Our
 −8/3 Z t  0 2/3 ‘event horizon’, at a cosmic time t, includes all
t t 0
ρr (t) = ρf e−Γφ t dt0 , points with which we will eventually communi-
tf 0 tf
(6.5) cate sending signals at t. The ‘instantaneous’
which, using the formula (at cosmic time t) radius of the ‘event horizon’ is
Z u Z ∞
X∞
up+k dt0
p−1 −x
x e dx = e −u
, de(t) = a(t) · (7.1)
p(p + 1) · · · (p + k) t a(t0)
0 k=0
(6.6) It is obvious, from this formula, that the ‘event
can be written as horizon’ is infinite for matter or radiation dom-
 
3 3 9 inated universe. For inflation, however, we ob-
ρ r = ρ Γ φ t 1 + Γφ t + (Γφ t)2 + · · · , tain a slowly varying ‘event horizon’ with radius
5 8 88
(6.7) de(t) = H −1 < ∞. Points, in our ‘event horizon’

9
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

at t, with which we can communicate sending outside the ‘event horizon’, goes as follows. A
signals at t, are eventually pulled away by the ‘comoving’ (present physical) scale `, at Tr , was
‘exponential’ expansion and we cease to be able equal to `(a(td )/a(t0)) = `(T0 /Tr ). Its magni-
to communicate with them again emitting signals tude at the end of inflation (t = tf ) was equal to
at later times. We say that these points (and the `(T0 /Tr )(a(tf )/a(td )) = `(T0 /Tr )(tf /td )(ν+2)/3ν
corresponding scales) crossed outside the ‘event ≡ `phys (tf ), where the potential V (φ) = λφν was
horizon’. The situation is very similar to that of assumed. The scale `, when it crossed outside
a black hole. Indeed, the exponentially expand- the inflationary horizon, was equal to H −1(φ` ).
ing (de Sitter) space is like a black hole turned We, thus, obtain
inside out. This means that we are inside and
the black hole surrounds us from all sides. Then, H −1(φ` )eN (φ` ) = `phys (tf ) . (7.4)
exactly as in a black hole, there are quantum
fluctuations of the ‘thermal type’ governed by Solving this equation, one can calculate φ` and,
the ‘Hawking temperature’ [26, 27] TH = H/2π. thus, N (φ` ) ≡ N` , the number of e-foldings the
It turns out [28, 29] that the quantum fluctua- scale ` suffered during inflation. In particular, for
tions of all massless fields (the inflaton is nearly our present horizon scale ` ≈ 2H0−1 ∼ 104 Mpc,
massless due to the ‘flatness’ of the potential) are it turns out that NH0 ≈ 50 − 60.
δφ = H/2π = TH . These fluctuations of φ lead Taking the potential V (φ) = λφ4 , Eqs.(4.9),
to energy density fluctuations δρ = V 0 (φ)δφ. As (7.3) and (7.4) give
the scale of this perturbations crosses outside the   √  3
‘event horizon’, they become [30] classical metric δρ 4 6π 1/2 φ`
= λ =
perturbations. ρ ` 5 MP
The evolution of these fluctuations outside
the ‘inflationary horizon’ is quite subtle and in- √  3/2
4 6π 1/2 N`
volved due to the gauge freedom in general rel- λ · (7.5)
5 π
ativity. However, there is a simple gauge invari-
ant quantity [31] ζ ≈ δρ/(ρ + p), which remains The measurements of COBE [2], (δρ/ρ)H0 ≈ 6 ×
constant outside the horizon. Thus, the density 10−5, then imply that λ ≈ 6 × 10−14 for NH0 ≈
fluctuation at any present physical (‘comoving’) 55. Thus, we see that the inflaton must be a
scale `, (δρ/ρ)` , when this scale crosses inside the very weakly coupled field. In nonsupersymmet-
post-inflationary particle horizon (p=0 at this in- ric GUTs, the inflaton is necessarily gauge sin-
stance) can be related to the value of ζ when the glet since otherwise radiative corrections will cer-
same scale crossed outside the inflationary ‘event tainly make it strongly coupled. This is, un-
horizon’ (symbolically at ` ∼ H −1 ). This latter doubtedly, not a very satisfactory situation since
value of ζ can be found using Eq.(4.5) and turns we are forced to introduce an otherwise unmo-
out to be tivated extra ad hoc very weakly coupled gauge
   0  singlet. In supersymmetric GUTs, however, the
δρ V (φ)H(φ)
ζ |`∼H −1 = = inflaton could be identified [32] with a conjugate
φ̇2 `∼H −1 2πφ̇2 `∼H −1
  pair of gauge nonsinglet fields φ, φ̄, already ex-
3
9H (φ) isting in the theory and causing the gauge sym-
=− · (7.2)
2πV 0 (φ) `∼H−1 metry breaking. Absence of strong radiative cor-
Taking into account an extra 2/5 factor from the rections from gauge interactions is guaranteed,
fact that the universe is matter dominated when in this case, by the mutual cancellation of the D
the scale ` re-enters the horizon, we obtain terms of these fields.
  √ The spectrum of density fluctuations which
δρ 16 6π V 3/2(φ` )
= · (7.3) emerge from inflation can also be analyzed. We
ρ ` 5 MP3 V 0 (φ` )
will again take the potential V (φ) = λφν . One
(ν+2)/2
The calculation of φ`, the value of the in- then finds that (δρ/ρ)` is proportional to φ`
flaton field when the ‘comoving’ scale ` crossed which, combined with the fact that N (φ` ) is pro-

10
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

portional to φ2` (see Eq.(4.9)), gives where ŝ(k̄) is a Gaussian random variable satis-
      ν+2 fying
δρ δρ N` 4
= . (7.6) 1
ρ ` ρ H0 N H0 < ŝ(k̄) >= 0 , < ŝ(k̄)ŝ(k̄0 ) >= δ(k̄ − k̄0) ,
k3
The scale ` divided by the size of our present hori- (8.5)
zon (≈ 104 Mpc) should equal exp(N` − NH0 ). and H is the amplitude of the perturbation when
This gives N` /NH0 = 1 + `n(`/104 )1/NH0 which its scale crosses inside the post-inflationary hori-
expanded around ` ≈ 104 Mpc and substituted zon. The latter can be seen as follows. A ‘comov-
in Eq.(7.6) yields ing’ (present physical) length ` crosses inside the
     αs post-inflationary horizon when a`/2π = H −1 =
δρ δρ ` a2/a0 which gives `/2π ≡ k −1 = a/a0 = ηH /2
= , (7.7)
ρ ` ρ H0 104 Mpc or kηH /2 = 1, where ηH is the ‘conformal’ time
at horizon crossing. This means that, at horizon
with αs = (ν + 2)/4NH0 . For ν = 4, αs ≈ 0.03
crossing, δk̄ (ηH ) = H ŝ(k̄). For scale invariant
and, thus, the density fluctuations are essentially
perturbations, the amplitude H is constant. The
scale independent.
gauge invariant perturbations of the scalar grav-
itational potential are given [33] by the Poisson’s
8. Density Fluctuations in ‘Matter’ equation,
We will now discuss the evolution of the primor- a2
dial density fluctuations after their scale enters Φ = −4πG ρδ (η) . (8.6)
k2 k̄
the post-inflationary horizon. To this end, we
From the Friedmann Eq.(8.2), we then obtain
introduce [33] the ‘conformal’ time, η, so that
the Robertson-Walker metric takes the form of a 3
Φ = − H ŝ(k̄) . (8.7)
conformally expanding Minkowski space: 2
The spectrum of the density perturbations
ds2 = −dt2 + a2(t) dr̄2 = a2(η) (−dη2 + dr̄2) ,
can be characterized by the correlation function
(8.1)
(x̄ is a ‘comoving’ 3-vector)
where r̄ is a ‘comoving’ 3-vector. The Hubble
parameter now takes the form H ≡ ȧ(t)/a(t) = ξ(r̄) ≡< δ̃ ∗ (x̄, η)δ̃(x̄ + r̄, η) > , (8.8)
a0 (η)/a2 (η) and the Friedmann Eq.(1.8) can be
rewritten as where Z
 2 δ̃(x̄, η) = d3kδk̄ (η)eik̄x̄ . (8.9)
1 a0 8πG
2
= ρ, (8.2)
a a 3 Substituting Eq.(8.4) in Eq.(8.8) and then using
where primes denote derivation with respect to Eq.(8.5), we obtain
the ‘conformal’ time η. The continuity Eq.(1.5) Z  4
3 −ik̄r̄ 2 kη 1
takes the form ρ0 = −3H̃(ρ + p) with H̃ = a0/a. ξ(r̄) = d ke H , (8.10)
2 k3
For a matter dominated universe, ρ ∝ a−3 which
gives a = (η/η0)2 and a0 /a = 2/η (η0 is the and the spectral function P (k, η) = 2H (η4 /16)k
present value of η). is proportional to k for H constant. We say that,
The Newtonian Eq.(2.7) can now be written in this case, the ‘spectral index’ n = 1 and we
in the form have a Harrison-Zeldovich [18] flat spectrum. In
a0 0 the general case, P ∝ k n with n = 1 − 2αs (see
δk̄00 (η) + δ (η) − 4πGρa2 δk̄ (η) = 0 , (8.3) Eq.(7.7)). For V (φ) = λφ4 , we get n ≈ 0.94.
a k̄
and the growing (Jeans unstable) mode δk̄ (η) is
proportional to η 2 and can be expressed [34] as 9. Temperature Fluctuations
 2
kη The density inhomogeneities produce tempera-
δk̄ (η) = H ŝ(k̄) , (8.4) ture fluctuations in the CBR. For angles θ ∼
>
2 o,
2

11
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

the dominant effect is the scalar Sachs-Wolfe [35] our present horizon scale during the inflationary
effect. Density perturbations on the ‘last scat- phase N`∼H −1 ≡ NQ ≈ 55.
0
tering surface’ cause scalar gravitational poten- There are also ‘tensor’ [36] fluctuations in
tial fluctuations, Φ, which, in turn, produce tem- the temperature of CBR. The quadrupole tensor
perature fluctuations in the CBR. The physical anisotropy is
reason is that regions with a deep gravitational  
potential will cause the photons to lose energy as δT V 1/2(φ` )
≈ 0.77 · (9.5)
they climb up the well and, thus, appear cooler. T Q−T MP2
For θ ∼<
2 o , the dominant effects are: i) Motion of The total quadrupole anisotropy is given by
the last scattering surface causing Doppler shifts,
  " 2  2 #1/2
and ii) Intrinsic fluctuations of the photon tem- δT δT δT
perature, Tγ , which are more difficult to calculate = + ,
T Q T Q−S T Q−T
since they depend on microphysics, the ionization
(9.6)
history, photon streaming and other effects.
and the ratio
The temperature fluctuations at an angle θ
 2
due to the scalar Sachs-Wolfe effect turn out [35] (δT /T )2Q−T MP V 0(φ` )
r= ≈ 0.27 · (9.7)
to be (δT /T )θ = −Φ` /3, ` being the ‘comoving’ (δT /T )Q−S
2 V (φ` )
scale on the ‘last scattering surface’ which sub-
tends the angle θ [ ` ≈ 100 h−1 (θ/degrees) Mpc ] For V (φ) = λφν , we obtain r ≈ 3.4 ν/NH  1,
and Φ` the corresponding scalar gravitational po- and the ‘tensor’ contribution to the temperature
tential fluctuations. From Eq.(8.7), we then ob- fluctuations of the CBR is negligible.
tain (δT /T )θ = (H /2)ŝ(k̄), which using Eq.(8.4)
gives the relation 10. Hybrid Inflation
   
δT 1 1 δρ
= δk̄ (ηH ) = · (9.1) 10.1 The non Supersymmetric Version
T θ 2 2 ρ `∼2πk−1
The most important disadvantage of the infla-
The COBE scale (present horizon) corresponds tionary scenarios described so far is that they
to θ ≈ 60 o . Eqs.(4.9), (7.3) and (9.1) give need extremely small coupling constants in order
    to reproduce the results of COBE [2]. This diffi-
δT V 3/2(φ` )
δρ ν+2
culty was overcome some years ago by Linde [37]
∝ 3 0
∝ ∝ N` 4 .
T ` ` ρ
MP V (φ`) who proposed, in the context of nonsupersym-
(9.2)
metric GUTs, an inflationary scenario known as
Analyzing the temperature fluctuations in spher-
hybrid inflation. The idea was to use two real
ical harmonics, the quadrupole anisotropy due to
scalar fields χ and σ instead of one that was nor-
the scalar Sachs-Wolfe effect can be obtained:
mally used. The field χ provides the vacuum en-
   1/2 3/2 ergy which drives inflation while σ is the slowly
δT 32π V (φ` )
= · (9.3) varying field during inflation. The main advan-
T Q−S 45 MP3 V 0 (φ` )
tage of this scenario is that it can reproduce the
For V (φ) = λφν , this becomes observed temperature fluctuations of the CBR
with ‘natural’ values of the parameters in con-
   1/2 ν+2
δT 32π λ1/2φ` 2 trast to previous realizations of inflation (like the
= =
T Q−S 45 νMP3 new [19] or chaotic [20] inflationary scenarios).
The potential utilized by Linde is
 1/2   ν+2
32π λ1/2 νMP2 4 ν+2  2
N`4
. (9.4) 2
2 χ2 λ2 χ 2 σ 2 m 2 σ 2
45 νMP3 4π V (χ, σ) = κ M − + + ,
4 4 2
Comparing this with COBE [2] measurements, (10.1)
(δT /T )Q ≈ 6.6 × 10−6, we obtain λ ≈ 6 × 10−14, where κ, λ are dimensionless positive coupling
for ν = 4, and number of e-foldings suffered by constants and M , m are mass parameters. The

12
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

vacua lie at hχi = ±2M , hσi = 0. Putting m=0, dard model (MSSM) plus standard model sin-
for the moment, we observe that the potential glets so that the successful predictions for αs,
possesses an exactly flat direction at χ = 0 with sin2θW are retained. The theory may also pos-
V (χ = 0, σ) = κ2 M 4. The mass squared of sess global symmetries. The breaking of G is
the field χ along this flat direction is given by achieved through the superpotential
m2χ = −κ2M 2 + λ2√ σ2 /2 and remains nonnega-
tive for σ ≥ σc = 2κM/λ. This means that, W = κS(−M 2 + φ̄φ), (10.3)
at χ = 0 and σ ≥ σc , we obtain a valley of min- where φ̄, φ is a conjugate pair of GS singlet left
ima with flat bottom. Reintroducing the mass handed superfields belonging to nontrivial repre-
parameter m in Eq.(10.1), we observe that this sentations of G and reduce its rank by their vevs
valley acquires a nonzero slope. A region of the and S is a gauge singlet left handed superfield.
universe, where χ and σ happen to be almost The coupling constant κ and the mass parameter
uniform with negligible kinetic energies and with M can be made positive by phase redefinitions.
values close to the bottom of the valley of min- This superpotential has the most general form
ima, follows this valley in its subsequent evolu- consistent with a U (1) R-symmetry under which
tion and undergoes inflation. W → eiθ W, S → eiθ S, φ̄φ → φ̄φ.
The quadrupole anisotropy of CBR produced The potential derived from the superpoten-
during this hybrid inflation can be estimated, us- tial W in Eq.(10.3) is
ing Eq.(9.3), to be
   1/2 V = κ2 | M 2 − φ̄φ |2 +κ2 | S |2 (| φ |2 + | φ̄ |2)
2 5
δT 16π λκ M
≈ . (10.2) +D − terms. (10.4)
T Q 45 MP3 m2
Restricting ourselves to the D flat direction φ̄∗ =
The COBE [2] result, (δT /T )Q ≈ 6.6 × 10−6, φ which contains the supersymmetric vacua and
can then be reproduced with M ≈ 2.86 × 1016 performing appropriate gauge and R- transfor-
GeV √ (the supersymmetric GUT vev) and m ≈ mations, we can bring S, φ̄, φ on the real axis,
1.3 κ λ ×1015 GeV ∼ 1012 GeV for κ, λ ∼ 10−2. √
i.e., S ≡ σ/ 2, φ̄ = φ ≡ χ/2, where σ, χ are
Inflation terminates abruptly at σ = σc and normalized real scalar fields. The potential then
is followed by a ‘waterfall’, i.e., a sudden entrance takes the form in Eq.(10.1) with κ = λ and m = 0
into an oscillatory phase about a global mini- and, thus, Linde’s potential for hybrid inflation
mum. Since the system can fall into either of the is almost obtainable from supersymmetric GUTs
two available global minima with equal proba- but without the mass term of σ which is, how-
bility, topological defects are copiously produced ever, of crucial importance since it provides the
if they are predicted by the particular particle slope of the valley of minima necessary for driv-
physics model one is considering. ing the inflaton towards the vacua.
One way to obtain a valley of minima useful
10.2 The Supersymmetric Version
for inflation is [39] to replace the renormalizable
The hybrid inflationary scenario is [38] ‘tailor trilinear superpotential term in Eq.(10.3) by the
made’ for application to supersymmetric GUTs next order nonrenormalizable coupling. Another
except that the mass of σ, m, is unacceptably way, which we will adopt here, is [40] to keep the
large for supersymmetry, where all scalar fields renormalizable superpotential in Eq.(10.3) and
acquire masses of order m3/2 ∼ 1 TeV (the grav- use the radiative corrections along the inflation-
itino mass) from soft supersymmetry breaking. ary valley (φ̄ = φ = 0 , S > Sc ≡ M ). In fact,
To see this, consider a supersymmetric GUT the breaking of supersymmetry by the ‘vacuum’
with a (semi-simple) gauge group G of rank ≥ 5 energy density κ2 M 4 along this valley causes a
with G → GS (the standard model gauge group) mass splitting in the supermultiplets φ̄, φ. This
at a scale M ∼ 1016 GeV. The spectrum of the results to the existence of important radiative
theory below M is assumed to coincide with the corrections on the inflationary valley. At one-
spectrum of the minimal supersymmetric stan- loop, and for S sufficiently larger than Sc , the

13
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

inflationary potential is given [40, 41] by hybrid inflationary scenario in a ‘natural’ way.
This means that a) there is no need of extremely
Veff (S) = κ2M 4 small coupling constants, b) W is the most gen-
   2 2  eral renormalizable superpotential which is al-
κ2 κ S 3 Sc4
1+ ln + − + · · · , lowed by the gauge and R- symmetries, c) su-
16π2 Λ2 2 12S 4
(10.5) persymmetry guarantees that the radiative cor-
where Λ is a suitable mass renormalization scale. rections do not invalidate inflation, but rather
From Eqs.(10.5) and (9.3), we find the cos- provide a slope along the inflationary trajectory
mic microwave quadrupole anisotropy: which drives the inflaton towards the supersym-
metric vacua, and d) supergravity corrections can
   1/2  2
δT NQ xQ M be negligible leaving inflation intact.
≈ 8π · (10.6)
T Q 45 yQ MP

Here NQ is the number of e-foldings suffered by Acknowledgments


our present horizon scale during inflation and
This work is supported by E.U. under TMR con-
yQ = xQ (1 − 7/(12x2Q ) + · · ·) with xQ = SQ /M ,
tract No. ERBFMRX–CT96–0090.
SQ being the value of the scalar field S when the
scale which evolved to the present horizon size
crossed outside the de Sitter (inflationary) hori- References
zon. Also, from Eq.(10.5), one finds [1] For a textbook treatment see e.g., S. Wein-
berg, Gravitation and Cosmology (Wiley, NY,
8π3/2 M
κ≈ p yQ · (10.7) 1972); E. W. Kolb and M. S. Turner, The Early
NQ MP Universe (Addison-Wesley, Redwood City, CA,
1990).
Inflation ends as S approaches Sc . Writing
S = xSc , x = 1 corresponds to the phase transi- [2] G. F. Smoot et al., Astrophys. J. Lett.
tion from G to GS which, as it turns out, more 396 (1992) L1; C. L. Bennett et al., Astrophys.
J. Lett. 464 (1996) 1.
or less coincides with the end of the inflationary
phase (this is checked by noting the amplitude [3] S. Burles and D. Tytler, Astrophys. Journal
of the quantities  and η in Eq.(4.7)). Indeed, 499 (1998) 699; ibid. 507 (1998) 732.
the 50 − 60 e-foldings needed for the inflationary [4] H. Georgi and S. Glashow, Phys. Rev. Lett. 32
scenario can be realized even with small values (1974) 438; J. G. Pati and A. Salam, Phys. Rev.
of xQ . For definiteness, we take xQ ≈ 2. From Lett. 31 (1973) 661.
COBE [2] one then obtains M ≈ 5.5 × 1015 GeV [5] D. A. Kirzhnits and A. D. Linde, Phys. Lett.
and κ ≈ 4.5 × 10−3 for NQ ≈ 56. Moreover, 42 B (1972) 471.
the primordial density fluctuation ‘spectral in- [6] M. Yoshimura, Phys. Rev. Lett. 41 (1978) 381;
dex’ n ' 0.98. We see that the relevant part of S. Dimopoulos and L. Susskind, Phys. Rev. D
inflation takes place at S ∼ 1016 GeV. An im- 18 (1978) 4500; D. Toussaint, S. Treiman, F.
portant consequence of this is [38, 41, 42] that Wilczek and A. Zee, Phys. Rev. D 19 (1979)
the supergravity corrections can be brought un- 1036; S. Weinberg, Phys. Rev. Lett. 42 (1979)
der control so as to leave inflation intact. 850; A. Y. Ignatiev, N. Y. Krosnikov, V. A.
After the end of inflation the system falls Kuzmin and A. N. Tavkhelidze, Phys. Lett.
76 B (1978) 326; J. Ellis, M. K. Gaillard and
towards the supersymmetric minima, oscillates
D. V. Nanopoulos, Phys. Lett. 80 B (1978) 360.
about them and eventually decays ‘reheating’ the
universe. The oscillating system (inflaton) con- [7] T. W. B. Kibble, J. Phys. A 9 (1976) 1387;
sists of the√two complex scalar fields S and θ = Phys. Reports 67 (1980) 183.
(δ φ̄ + δφ)/ 2, where δ φ̄ = φ̄ − M , δφ = φ − M , [8] G.’t Hooft, Nucl. Phys. B 79 (1974) 276; A.

with mass minf l = 2κM . Polyakov, JETP Lett. 20 (1974) 194.
In conclusion, it is important to note that [9] H. B. Nielsen and P. Olesen, Nucl. Phys. B 61
the superpotential W in Eq.(10.3) leads to the (1973) 45.

14
Corfu Summer Institute on Elementary Particle Physics, 1998 G. Lazarides

[10] Ya. B. Zeldovich, I. Yu. Kobzarev and L. B. [28] T. Bunch and P. C. W. Davies, Proc. Roy. Soc.
Okun, JETP (Sov. Phys.) 40 (1975) 1. London A 360 (1978) 117.
[11] J. P. Preskill, Phys. Rev. Lett. 43 (1979) 1365. [29] A. Vilenkin and L. Ford, Phys. Rev. D
26 (1982) 1231; A. D. Linde, Phys. Lett.
[12] A. H. Guth, Phys. Rev. D 23 (1981) 347.
116 B (1982) 335; A. A. Starobinsky, Phys.
[13] For a textbook introduction see e.g., A. D. Lett. 117 B (1982) 175.
Linde, Particle Physics and Inflationary Cos-
[30] W. Fischler, B. Ratra and L. Susskind, Nucl.
mology (Harwood Academic Publishers, Lon-
Phys. B 259 (1985) 730.
don, 1990).
[31] J. M. Bardeen, P. J. Steinhardt and M. S.
[14] Ya. B. Zeldovich, Mon. Not. Roy. Astr. Soc.
Turner, Phys. Rev. D 28 (1983) 679; for a
192 (1980) 663.
review see e.g.,V. F. Mukhanov, H. A. Feld-
[15] For a pedagogical discussion see e.g., P. J. E. man and R. H. Brandenberger, Phys. Reports
Peebles, The Large-scale Structure of the Uni- 215 (1992) 203.
verse (Princeton Univ. Press, Princeton, 1980);
[32] G. Lazarides and Q. Shafi, Phys. Lett. B 308
G. Efstathiou, in The Physics of the Early Uni-
(1993) 17.
verse, eds. J. A. Peacock, A. F. Heavens and A.
T. Davies (Adam-Higler, Bristol, 1990). [33] J. M. Bardeen, Phys. Rev. D 22 (1980) 1882;
H. Kodama and M. Sasaki, Prog. Theor. Phys.
[16] See e.g., P. J. E. Peebles, D. N. Schramm, E. L.
72 (1984) 1.
Turner and R. G. Kron, Nature 352 (1991) 769.
[34] L. F. Abbott and R. K. Schaefer, Astrophys.
[17] V. L. Ginzburg, Sov. Phys. -Solid State
Journal 308 (1986) 546; R. K. Schaefer, Phys.
2 (1960) 1824. Rev. D 6 (1987) 2075.
[18] E. R. Harrison, Phys. Rev. D 1 (1970) 2726; [35] R. K. Sachs and A. M. Wolfe, Astrophys. Jour-
Ya. B. Zeldovich, Mon. Not. Roy. Astr. Soc. nal 147 (1967) 73.
160 (1972) 1.
[36] V. A. Rubakov, M. Sazhin and A. Veryaskin,
[19] A. D. Linde, Phys. Lett. 108 B (1982) 389; A. Phys. Lett. 115 B (1982) 189; R. Fabbri and
Albrecht and P. Steinhardt, Phys. Rev. Lett. 48 M. Pollock, Phys. Lett. 125 B (1983) 445; B.
(1982) 1220. Allen, Phys. Rev. D 37 (1988) 2078; L. Abbott
[20] A. D. Linde, JETP Lett. 38 (1983) 149; Phys. and M. Wise, Nucl. Phys. B 244 (1984) 541; M.
Lett. 129 B (1983) 177. White, Phys. Rev. D 46 (1992) 4198.
[21] R. J. Scherrer and M. S. Turner, Phys. Rev. D [37] A. D. Linde, Phys. Lett. B 259 (1991) 38; Phys.
31 (1985) 681. Rev. D 49 (1994) 748.

[22] M. Yu. Khlopov and A. D. Linde, Phys. Lett. [38] E. J. Copeland, A. R. Liddle, D. H. Lyth, E. D.
138 B (1984) 265; J. Ellis, J. E. Kim and Stewart and D. Wands, Phys. Rev. D 49 (1994)
D. Nanopoulos, Phys. Lett. 145 B (1984) 181; 6410.
M. Kawasaki and T. Moroi, Prog. Theor. Phys. [39] G. Lazarides and C. Panagiotakopoulos, Phys.
93 (1995) 879. Rev. D 52 (1995) R559.
[23] A. Albrecht, P. J. Steinhardt, M. S. Turner and [40] G. Dvali, Q. Shafi and R. K. Schaefer, Phys.
F. Wilczek, Phys. Rev. Lett. 48 (1982) 1437; Rev. Lett. 73 (1994) 1886.
L. Abbott, E. Farhi and M. Wise, Phys. Lett.
[41] G. Lazarides, R. K. Schaefer and Q. Shafi, Phys.
117 B (1982) 29; A. Dolgov and A. D. Linde,
Rev. D 56 (1997) 1324.
Phys. Lett. 116 B (1982) 329.
[42] G. Lazarides and N. Tetradis, Phys. Rev. D 58
[24] P. J. Steinhardt and M. S. Turner, Phys. Rev.
(1998) 123502; E. D. Stewart, Phys. Rev. D 51
D 29 (1984) 2162.
(1995) 6847.
[25] M. S. Turner, Phys. Rev. D 28 (1983) 1243.
[26] S. W. Hawking, Nature 248 (1974) 30; Com-
mun. Math. Phys. 43 (1975) 199.
[27] G. W. Gibbons and S. W. Hawking, Phys. Rev.
D 15 (1977) 2752.

15

Anda mungkin juga menyukai