Cosmic Strings

Kevin Roberts
December 13, 2010

Abstract
Cosmic strings are similar to other one-dimensional topological de-
fects familiar from condensed matter systems such as vortex lines in
type II superconductors. The goal of this paper is to provide a brief
introduction to what cosmic strings are and how they come about in
the early stages of the universe highlighting similarities to defects from
condensed matter systems when possible. After will be a discussion
of their possible observational consequences and their importance to
cosmology; especially gravitational lensing and the effect on their cos-
mic microwave background (CMB). I will conclude by attempting to
give a brief overview of the current state of cosmic string research.

1
1 Introduction
The term “cosmic strings” refers to 1D topological defects formed in sym-
metry breaking phase transitions in the early universe. Such defects are
familiar from systems studied in condensed matter such as disclination lines
in nematic liquid crystals, vortex lines in liquid 4 He, and flux tubes in type
II superconductors. Interestingly, analogous defects occur in field theories
normally relegated to high energy physics and with similar dynamics.
The idea that strings could be stable and have observational consequences
was first proposed by Tom Kibble in the 1970’s [1]. 2D domain walls and
0D monopoles are also possible but are ruled out by experience. Due to
gravitational effects domain walls would ruin the isotropy of the cosmic mi-
crowave backround (CMB) [1][2]. Stable, ultra-heavy monopoles are generic
predictions of grand unified theories (GUTs). Once formed, they would soon
dominate the energy density of the Universe. This is known as the “monopole
problem” because we don’t see any. Cosmic inflation was developed, in part,
to solve the monopole problem by reducing the monopole density to unde-
tectable levels. Inflation should also then reduce the density of cosmic strings.
However, it is possible to reconcile these theories by adjusting the coupling
between the inflation field and the field responsible for the cosmic strings so
that string formation occurs during the late stages of inflation. For more
details see the references given in [3, sec. 5.2].
The study of cosmic strings is a very rich field and encompasses aspects of
high energy physics, cosmology, numerical simulation, and the experimental
efforts of observational astronomers. I hope to tell a complete, if simplified,
version of how strings come about, how they evolve, and how they may affect
our Universe today.
Please note that in this paper I will be using natural units in which
h̄ = c = kB = 1.

2 String solutions in field theories

Much of the discussion in this section follows closely the terrific review article
by Hindmarsh and Kibble [3]. In this section I will attempt to provide simple
examples of string solutions in field theories and show how they are similar
to more familiar condensed matter analogs.

2
2.1 Global strings
The simplest example of a field theory with string solutions is that of a
complex scalar field φ(x) with a Lagrangian density
1 1
L = ∂ µ φ∗ ∂µ φ − V (φ) V (φ) = λ(|φ|2 − η 2 )2
2 2
This Lagrangian posesses a global U(1) symmetry φ → φeiα with α a con-
stant.
√ It also exhibits a double well potential with a ground state φ =
(η/ 2)eiα0 where α0 is a constant. This ground state is not unvariant un-
der the global U(1) symmetry. The equation of motion following from this
Lagrangian is
1
∂µ ∂ µ φ − λη 2 φ + λ|φ|2 φ = 0.
2
The mass of the scalar particle in the symmetry-breaking vacuum is given by
m2s = λη 2 . In natural units mass has dimensions of L-1 and m−1
s is analogous
to the correlation length in superconductors as it determines the scale of
fluctuations in φ.
We now insert a cylindrically symmetric trial solution φ = √η2 f (ms ρ)einϕ
where (ρ, ϕ, z) are cylindrical coordinates and n an integer. The equation of
motion reduces to a single non-linear ODE
1 n2 1
f 00 + f 0 − 2 f − (f 2 − 1)f = 0
ξ ξ 2
where ξ = ms ρ. To preserve the continuity of φ at the origin we must have
that f → 0 as ξ → 0. We also need f → 1 at infinity so that φ approaches
its ground state value. If we write f = 1 − δf and plug this into our ODE
we see that δf ∼ n2 /ξ 2 at large ξ. A numerical solution for f (ξ) is shown
below.
Notice that if n = 0 then f = 1 is a valid solution and we lose this behavior
at the origin. These solutions are closely tied to vortices in superfuid 4 He
where n is the topological charge of the vortex and tells us the number of
times the phase winds around. n = 0 corresponds to there being no vortex.
If we now examine the energy density

E = |φ̇|2 + |∇φ|2 + V (φ)

n2 2 1 2
∝ (f 0 )2 + f + (f − 1)f 2
ξ2 2

3
 Figure 1: f (ξ)

we see that it is infinite because it has a ξ −2 tail at large ξ due to the the
angular derivative part of the gradient. Inside a cylinder of radius R >> m2s
the energy per unit length is approximately πn2 η 2 ln(ms R).

2.2 Local strings

Let us now add a guage field Aµ . Our Lagrangian density is then
1
L = − Fµν F µν + |Dµ φ|2 − V (φ)
4
where Dµ = ∂µ + ieAµ and Fµν = ∂µ Aν − ∂νAµ .
This is the Abelian Higgs model. This Lagrangian has a local U(1) sym-
metry and is invariant under the transformation
1
φ → φieΛ(x) and Aµ → Aµ − ∂µ Λ(x)
e
The equations of motion for this Lagrangian are
1
Dµ Dµ φ + λ(|φ|2 − η 2 )φ = 0
2
∂νF µν + ie(φ∗ Dµ φ − Dµ φ∗ φ) = 0
The scalar Higgs particle, once again, has a mass m2s = λη 2 in the symmetry
breaking vacuum. The vector field also aquires a mass of mv = eη.

4
 Now let us take the radial gauge Aρ = 0 and try the cylindrically sym-
metric solutions
η n
φ = √ f (mv ρ)einϕ Ai = a(mv ρ)ϕ̂i
2 eρ
The resulting coupled differential equations do not have an analytic solution
but their asymptotic behavior is as follows:
f0 ξ |n|

√ as ξ → 0
f' −1/2
1 − f1 ξ exp(− βξ) as ξ → ∞
(
|n|f02
a0 ξ 2 − 4(|n|+1) ξ 2|n|+2 as ξ → 0
a' 1/2
1 − a1 ξ exp(−ξ) as ξ → ∞
Here, ξ √= mv ρ and β = λ/e2 = (ms /mv )2 . In the case where β > 4,
−1/2
ξ exp(− βξ) is replaced by ξ −1 exp(−2ξ). Notice that everything is much
more localized on the string. In fact, it can be shown that the energy density
is µ = πη 2 (β) where  goes as log(β) for β > 1 [6][7] and (1) = 1 [5].
The energy density here is finite because the gradients have been replaced
by covariant derivatives so that the angular derivative part now goes down
faster than ξ −2 .
Now consider a circle C of very large radius centered on the string encir-
cling a surface S. On a circle whose radius tends towards infinity Dµ φ must
vanish. We can use this to calculate the magnetic flux which passes through
the string.
Z Z
2πn
B · n̂d x = Ai dxi =
2
S C e
The magnetic flux is quantized much like that of vortices in superconduc-
tors.
There is another analogy that can be made. Vortices in the Abelian
Higgs model are stable for any n if β < 1. For β > 1 a vortex with |n| > 1
is unstable to splitting into n vortices with charge n = 1. This can be
interpreted in much the same way as distinguishing between type I and II
superconductors. There are two forces at work here. The scalar field would
like to push the vortices on top of each other since it wants to minimize the
area where φ = 0. However, the gauge field wants to push the vortices apart
since the magnetic field lines want to spread out. The long-range force, i.e.,
the field with the lower mass, is the one that dominates. So, the β parameter
is somewhat analogous to the Landau-Ginzburg parameter κ.

5
3 Phase transitions in the early universe
Now that we have seen what strings solutions look like, let’s look at how they
might form in the early universe. Let us begin with the simple complex scalar
field Lagrangian from section 2.1. At a finite temperature the expectation
value of φ must be obtained by minimizing the free energy or, equivalently,
the effective temperature dependent potential. If we assume the coupling
constant λ is small then the leading temperature dependent terms at high T
can be found by calculating the one-loop diagrams [4]. Our potential then
becomes
1 1 1
V (φ) = λ(|φ|2 − η 2 ) + λT 2 |φ|2 .
2 2 6
√
If we now define Tc ≡ 3η the potential can be rewritten as
1 1
V (φ) = λ[|φ|2 − η 2 (1 − T 2 /Tc2 )]2 + λη 4
2 2
which now looks like the homogenous part of the Ginzburg-Landau free en-
ergy. Above Tc the potential has a singular minimum at φ = 0. Below Tc , φ
will aquire a vacuum expectation value such that |φ|2 = η 2 (1 − T 2 /Tc2 ). Also
notice that in section 2 we saw that a string’s energy density µ ∼ η 2 . This
means that µ ∼ Tc2 .
This simple model illustrates a second order transition. In this type of
transition, when the temperature falls below critical, the field will assume
a vacuum expectation value everywhere at about the same time. Due to
causality reasons, the value of the phase at widely different postions must be
uncorrelated. This differs somewhat from the behavior in superconductors
where one normally thinks of the correlation length extending to infinity at
the critical temperature. Strings will then form where the phase winds a
multiple of 2π around a loop.
Whether or not higher order terms render the phase transition of first-
order is a complicated issue. If the phase transition is of first-order it would
then proceed by bubble nucleation. The field will roll off the potential hill at
different places at different times. One expects the phase of the field in each
bubble to be more or less a random variable. When bubbles meet they tend
to interpolate between their respective values of φ at the boundary. Regions
which get trapped in a loop where the phase winds around by a multiple of
2π will form strings.

6
 Figure 2: The formation
of global strings. Bub-
bles nucleate (a), collide
(b), and merge (c). If the
phases of the field in the
bubbles are widely dis-
tributed it is possible to
trap a region of normal
phase in between them
and form a string. [3, sec.
5.2]

How early in the universe do these transitions take place? One can find
a relation between the time and temperature in the early universe when it
was dominated by relativistic matter
r
45 Mpl 2.42
2
T t= = 1/2 MeV2 s
πg∗ 4π g∗

where g∗ (which we take to be of order ∼ 100 during the relevant time period)
is the effective number of spin states and Mpl is the Planck mass G−1/2 =
1.22x1028 eV [8]. In the electroweak unified theory there is a phase transition
a temperature of order 100 GeV (the rest mass of the W and Z gauge bosons).
This corresponds to a time of 10−5 s after the big bang. In most GUTs, a

7
phase transition occurs at around 1015 GeV or 1016 GeV corresponding to a
time between 10−39 s and 10−37 s. This is, of course, assuming our theories of
particle physics work at such high temperatures.

4 Observational consequences
Cosmic strings are interesting theoretical objects. The question remains as
to whether or not they actually exist. I will briefly sketch a few of the ways
in which one might observe a cosmic string along with results from a few
more notable searches.

4.1 Gravitational lensing

Space-time around a string is coned shaped as though a wedge of angle δ were
cut out. The deficit angle is related
to the energy density µ by

δ = 8πGµ.

For strings formed at the elec-

troweak scale this is not observable
but for strings formed at GUT en-
ergy scales this corresponds to an
angle of 5.18x10−5 ∼ 10−6 radians
which is. Figure 3: An illustration of the
The conical shaped space time cone-shaped space time and resulting
around a cosmic string lends itself to geodesics around a string [12].
detection by observing gravitational
lensing. What one would look for would be a double image of a distant
object; one on each side of the string.
In the early part of this decade, an object appearing to be a lensed galaxy
called CSL-1 was proposed as a candidate for string lensing [9][10]. The
two galaxies observed have nearly the same morphologies, magnitudes, and
spectra. However, closer inspection by the Hubble telescope in 2006 revealed
them to be two very similar galaxies [11]. The principle axes of the galaxies
are misaligned which is not an expected effect of any sort of gravitational
lensing.

8
 Then there is the interesting case of Q0957+561, a system of two quasar
images. The images fluctuate in brightness with a delay between them of 417
days. It is a well known lensed system whose delay has been used to mea-
sure the Hubble constant. The source of the lensing is a foreground galaxy
whose image is clearly visible between the two galaxies. However, in 2004, a
group led by Robert Schild reported observing a component which fluctuated
synchronously, with no delay, dur-
ing the period between September
2004 and July 2005 [13]. They pro-
posed a scenario where the reason
for this synchronous fluctuation is an
oscillating cosmic string loop which
passed between us and the lensing
object. While this explanation may
sound far fectched, other explana-
tions involving multiple microlens-
ing stars and undetected, massive bi-
nary star systems seem equally un-
Figure 4: How correlated are these likely. There is also the question of
curves? [13] how statistically significant the cor-
relation between the brightness fluc-
tuations really are; an analysis that has yet to be done.

4.2 Effects on the CMB

Measuring the inhomogeneity of the cosmic microwave background was once
hoped to provide the most definitive test of the existence of cosmic strings.
The effect of a moving cosmic string is to redshift the light in front of it and
blueshift the light behind it an amount proportional to δ and the velocity
of the string perpindicular to the line of sight. This will show up when
measuring the spectrum of the CMB. To predict the effect on the CMB one
needs to know what the distribution in space and velocity of the strings will
be. This inevitably involves a lot of numerical work. The Relativity and
Gravitation research group at Cambridge maintains a webpage with plenty
of movies of simulated string dynamics and evolution [12].
It became very difficult to produce simulations which would match the
large-scale structure and CMB simultaneously. Meanwhile, the theory that
the inhomogeneities were caused by the inflation of primordial quantum den-

9
 Figure 5: Fits of data to measured CMB power spectrum [14]

sity fluctuations was fitting the data much better [14]. This ultimately led
to the decline in interest of cosmic strings in the early 2000’s.

4.3 Other searches

There are many other proposed methods for searching for cosmic strings. One
important way is to test cosmic string theory predictions against observed
density perturbations. This leads to similar sorts of analyses that one gets
into when measuring the CMB. Another way is the detection of gravitational
waves caused by strings, perhaps by LIGO [15]. There are also theories that
moving cosmic strings may be the origins of the highest energy cosmic rays
(watch some of the movies on the Cambridge website and notice particles
‘evaporating’ from the cusps of the strings) but this seems unlikely [3, sec.
6.5].

5 Current state of cosmic string research

10
 Interest in cosmic strings waned in
the early part of this decade when it
was discovered that they could not
play a critical role in structure for-
mation.
However, in 2004, interest in cos-
mic strings exploded. The new in-
terest in cosmic strings is due to two
reasons: one observational, the other
theoretical. On the observation side,
there was a lot of excitement about
the possible discovery of a cosmic
string due to CSL-1 and the intrigu-
Figure 6: Number of article about cos- ing case of the anomalous time delay
mic strings and cosmic superstrings on in Q0957+561 A, B. On the theoret-
arxiv.org ical side, there was increased inter-
est due to the realization that string
theory could produce stable cosmic strings of the appropriate energy scale.

6 Conclusion and future steps

Cosmic strings have been studied in one form or another since the 1970’s
and as such has become a very rich and detailed field of study. This paper
can only be a short introduction at best. Most notable was an ommission
of any mention of string dynamics and how the numerical simulations are
done. This is still an active area of research. The issue of string dynamics
and interactions is extremely interesting as a study of how one may go about
describing a theory of vortex matter. I refer you section 3 of Hindmarsh and
Kibble’s review article for more details.
The prospects for the study of cosmic strings beyond being a toy of the-
orists seems doubtful. The few supposed detections are all questionable at
best and completely debunked at worst. The prospect of a confirmed detec-
tion of single strings moving near the speed of light seems dim and it seems
likely that strings suffered the same fate as monopoles, inflation reducing
their density to undetectable levels (if they ever existed at all). Glimmers
of hope come from the expectation of more accurate measurements of den-
sity perturbations coming from Dark Energy Survey and the Large Synoptic

11
Survey Telescope. Better measurements of CMB polarization may also hold
valuable information. Perhaps with increasingly better measurements of as-
tronomical data we will be able to place better limits on just what kind of
contribution defects have on CMB inhomogeneities and large-scale structue.

References
[1] TWB Kibble 1976 J. Phys. A:Math. Gen. 9 1387
[2] Zel’dovich Ya B, Kobzarev I Yu and Okun LB 1974 Zh. Eksp. Teor. Fiz.
67 3
[3] M B Hindmarsh and T W B Kibble 1995 Rep. Prog. Phys. 58 477
[4] Weinberg S 1974 Phys. Rev. D 9 3357; Dolan L and Jackiw R 1974
Phys. Rev. D 9 3320; Bernard C 1974 Phys. Rev. D 9 3312
[5] Bogomol’nyi E B 1976 Sov. J. Nucl. Phys. 24 449 [Yad. Fiz. 24 861
(1976)]
[6] Jacobs L and Rebbi C 1979 Phys. Rev. B 19 4486
[7] Hill C T, Hodges H M and Turner M S 1988 Phys. Rev. D 37 263
[8] Kolb E Wand Turner M S 1990 The Early Universe (Redwood City:
Addison-Wesley)
[9] M. Sazhin, et al., Mon. Not. Roy. Astron. Soc. 343, 353 (2003)
[10] M. Sazhin, et al., Astrophys. J. 636, L5 (2006)
[11] Agol E, Craig H, and Richard M P 2006 Phys. Rev. D 73 087302
[12] http://www.damtp.cam.ac.uk/research/gr/public/cs home.html
[13] R. Schild, I. S. Masnyak, B. I. Hnatyk and V. I. Zhdanov 2004 Astronomy
and Astrophysics 422 477
[14] F.R. Bouchet, P. Peter, A. Riazuelo & M. Sakellariadou 2002 Phys. Rev.
D 65 021301.
[15] arXiv:0904.4718v2 [astro-ph.CO]

12