PHYSICAL REVIEW D 82, 025022 (2010)

Fate of the false monopoles: Induced vacuum decay

Brijesh Kumar,1,2,* M. B. Paranjape,2,† and U. A. Yajnik1,2,3,‡
1
Physics Department, Indian Institute of Technology Bombay, Mumbai, 400076, India
2
Groupe de physique des particules, Département de physique, Université de Montréal, Case Postale 6128, succursale Centre-ville,
Montréal, Québec, Canada, H3C 3J7
3
Department of Physics, Ernest Rutherford Physics Building, McGill University, 3600 rue University,
Montréal, Québec, Canada, H3A 2T5
(Received 3 June 2010; published 30 July 2010)
We study a gauge theory model where there is an intermediate symmetry breaking to a metastable
vacuum that breaks a simple gauge group to a Uð1Þ factor. Such a model admits the existence of
metastable magnetic monopoles, which we dub false monopoles. We prove the existence of these
monopoles in the thin-wall approximation. We determine the instantons for the collective coordinate
that corresponds to the radius of the monopole wall and we calculate the semiclassical tunneling rate for
the decay of these monopoles. The monopole decay consequently triggers the decay of the false vacuum.
As the monopole mass is increased, we find an enhanced rate of decay of the false vacuum relative to the
celebrated homogeneous tunneling rate due to S. R. Coleman [Subnuclear series 13, 297 (1977).].

DOI: 10.1103/PhysRevD.82.025022 PACS numbers: 12.60.Jv, 11.27.+d

obtained is generically metastable [13] and its phenome-

I. INTRODUCTION
nological viability depends on the tunneling rate being
Semiclassical solutions with topologically nontrivial sufficiently slow.
boundary conditions in relativistic field theory [1–3] have A change in phase due to metastable topological objects
the interesting property that they interpolate between two is a generalization of the following better known mecha-
or more alternative translationally invariant vacua of the nism. When the effective potential of the theory possesses
theory. For instance the exterior of a monopole or a vortex several local minima, all but the lowest minimum are
solution is a phase of broken symmetry, while the interior quantum mechanically unstable. The so-called false vacua
of the object generically contains a limited region of un- are then liable to decay, even in the absence of topological
broken symmetry (for more details and lucid expositions objects, according to a rate given by a WKB-like formula
see [4,5]). Most of the commonly studied solutions are studied earlier in [14] and provided an elegant and lucid
topologically nontrivial; however, nontrivial boundary footing by Coleman [15,16]. The cases studied there con-
conditions are not a guarantee of dynamical stability. In cerned a transition between two translationally invariant
[6], for example, a large number of such solutions are vacua. The generic scenario of decay consists of sponta-
constructed in gauge field theories which are generically neous formation of a small bubble of true vacuum, which
metastable. The Skyrmion is also a classic example of a can then start growing by semiclassical evolution. In
topologically nontrivial configuration that is unstable with- Minkowski space, the formation of one such bubble is
out the addition of a fourth order Skyrme term [7,8]. All of sufficient to convert the phase of the system to the true
the classically stable solutions (allowing for quantum vacuum. In the context of an expanding Universe, conver-
metastabilty) are nontrivial time independent local minima sion of the entire Universe to the true vacuum would
of the effective action of the theory. require formation of a sufficiently large number of such
The metastability of such solutions can be of significant bubbles at an adequate rate.
interest. The implied decay of the object would be accom- The existence of topological objects may provide addi-
panied by the change in phase of the system as a whole. In tional sources of metastability. Phase transitions seeded by
the context of cosmology this may imply a change in the topological solutions were studied early in the works of
cosmic history and determine the abundance of relic ob- [17–20]. An essential aspect of these studies is precisely
jects. On a more formal footing the question of metastabil- the observations that there exist solutions with nontrivial
ity of vacua has gained considerable interest in the context boundary conditions which interpolate between two dis-
of supersymmetric field theories [9] where a nonsupersym- tinct minima of the effective potential. The importance of
metric phase is required on phenomenological grounds but this alternative route to decay arises from the fact that it can
such a phase is necessarily metastable on theoretical be much more rapid than the spontaneous decay of a
grounds [11,12]. In string cosmology the de Sitter solution translationally invariant vacuum. Indeed, for some values
of the parameters the decay induced by topological objects
*[email protected] may require no tunneling and therefore would be very
†
[email protected] prompt in a context where the parameters are changing
‡
[email protected] adiabatically, as for instance in the early Universe.

1550-7998= 2010=82(2)=025022(10) 025022-1 Ó 2010 The American Physical Society

BRIJESH KUMAR, M. B. PARANJAPE, AND U. A. YAJNIK PHYSICAL REVIEW D 82, 025022 (2010)
Obtaining a general formula characterizing this kind of tunneling to a new configuration of a much larger radius
vacuum decay has been rather elusive although the ideas and we determine the existence of the instanton for this
have been adequately explicated in [17–20]. More recently, tunneling within the same thin-wall approximation. In
the relevance of the mechanism has been demonstrated in Sec. V we determine the Euclidean action for this instan-
specific examples, in [21] for the mediating sector of a ton, the so-called bounce B which determines the tunneling
hidden sector scenario of supersymmetry breaking and in rate for the appearance of the large radius unstable mono-
[22] in a grand unified theory model with O’Raifeartaigh pole. In Sec. VI we relate our findings to a previous study
type direct supersymmetry breaking. In this paper we of classical monopole instability in supersymmetric grand
explore a model that is amenable to an analytical treatment unified theory models. In Sec. VII we discuss our results
within the techniques developed in [23]. In doing so we and compare our tunneling rate formula with that of the
provide a transparent model in which the generic expecta- homogeneous bubble formation case without monopoles.
tions raised in [17–20] can be realized and a specific We show that in addition to our tunneling rate being
formula can be derived. significantly faster, it also indicates a regime in which
We construct an SUð2Þ gauge model with a triplet scalar the monopoles become unstable, hence showing that the
field with two possible translationally invariant vacua, one putative nontrivial vacuum indicated by the effective po-
with SUð2Þ broken to Uð1Þ and the other with the original tential is in fact unstable.
gauge symmetry intact. The former phase permits the
existence of monopoles. By appropriate choice of potential II. UNSTABLE MONOPOLES IN A FALSE VACUUM
for the triplet it can be arranged that the phase of unbroken
symmetry is lower in energy and represents the true vac- Consider an SUð2Þ gauge theory with a triplet scalar
uum of the theory. The monopoles interpolate between the field
with the Lagrangian density given by
true vacuum and the false vacuum. For a wide range of the L ¼
14F
a
Fa þ 12ðD
a ÞðD
a Þ
Vð
a
a Þ; (1)
parameters, these monopoles are in fact classically stable.
In previous work [17,18] the dissociation of such mono- where
poles was considered, varying the parameters of the theory a ¼ @ Aa
@ Aa þ eabc Ab Ac ;
F (2)
     
to critical values where the monopoles were classically
unstable due to infinite dilation. This can occur, for ex- and
ample, in the early Universe where the high temperature
phase prefers one vacuum in which the system starts, but D
a ¼ @
a þ eabc Ab
c : (3)
with adiabatic reduction in temperature, a different phase The potential we use is a polynomial of order 6 in
and
becomes more favorable. The Universe is then liable to may conveniently be written as
simply roll over, by classical evolution, to the true vacuum.
It was, however, overlooked that these monopoles are in Vð
Þ ¼ 
2 ð
2
a2 Þ2 þ 2
2
; (4)
fact unstable due to quantum tunneling well before the where  is defined so that the potential vanishes at the
parameters reach their critical values. We dub such mono- metastable vacua. The vacuum energy density difference is
poles false monopoles. Working in the thin-wall limit for then equal to . Such a potential was numerically analyzed
the monopoles [17], we show that such monopoles undergo in [24] as a toy model for the dissociation of monopoles.
quantum tunneling to larger monopoles, which are then Here we obtain explicit analytical formulas for the quan-
classically unstable by expanding indefinitely, conse- tum tunneling decay of the monopoles. The potential has a
quently converting all space to the true vacuum, the phase minimum at
T
¼ 0 which for  ¼ 0 is degenerate with
of unbroken SUð2Þ symmetry. Further, the formula we the manifold of vacua at
T
¼ a2 . When we set  Þ 0,
derive also recovers the regime of parameter space, within we get a manifold of degenerate metastable vacua at
the thin-wall monopole limit, where no tunneling is re-
T
¼ 2 (where the exact value of the vacuum expec-
quired for the decay but the monopole is simply classically tation value, , is calculable and satisfies   a for small
unstable as previously treated [17,18]. ), and the minimum at
¼ 0 becomes the true vacuum.
The rest of the paper is organized as follows. In Sec. II A plot of the potential for small  as a function of one of
we specify the model under consideration and the mono- the components of
is shown in Fig. 1. A supersymmetry
pole ansatz along with the equations of motion. In Sec. III breaking model [25] containing monopoles and a scalar
we delineate the conditions in which there should exist a potential similar to the one given in Eq. (4) was studied
metastable monopole solution with a large radius and a thin in [22].
wall. We find the thin-wall monopole solutions and also The manifold of vacua at
T
¼ 2 is topologically an
justify their existence. In Sec. IV we use the thin-wall S , and as spatial infinity is topologically also S2 ,
2

approximation which permits a treatment of the solution the appropriate homotopy group of the manifold of the
in terms of a single collective coordinate, the radius R of vacua of the symmetry breaking SUð2Þ ! Uð1Þ is
the thin wall. We argue that the monopole is unstable to
2 ðSUð2Þ=Uð1ÞÞ which is Z. This suggests the existence

025022-2
FATE OF THE FALSE MONOPOLES: INDUCED VACUUM DECAY PHYSICAL REVIEW D 82, 025022 (2010)
III. THIN-WALLED MONOPOLES
When the difference between the false and true vacuum
energy densities  is small, the monopole can be treated as
a thin shell, the so-called thin-wall approximation. Within
Potential V(φ) -->

this approximation, the monopole can be divided into three

regions as shown in Fig. 2. There is a region of essentially
true vacuum extending from r ¼ 0 up to a radius R. At r ¼
R, there is a thin shell of thickness in which the field
value changes exponentially from the true vacuum to the
φ −−> false vacuum. Outside this shell the monopole is essentially
−η η
in the false vacuum, and so we have
−ε

h  0; K  1
r<R
;
2
FIG. 1 (color online). The potential Vð
Þ for  Þ 0 as a (9)
function of one of the components of the field
, shifted by h  ; K  0 r>Rþ ;
2
an additive constant so that
¼  has vanishing V and the true
vacuum has V ¼
.
0 < h < ; 0 < K < 1 R
rRþ ;
2 2
of topologically nontrivial solutions of the monopole type where is a length corresponding to the mass scale of the
which are classically stable. The presence of the global symmetry breaking. As we shall see in Sec. IV, describing
minimum at
¼ 0 allows for the possibility that the the monopole in this way allows us to study the dynamics
monopole solution, although topologically nontrivial, in terms of just one collective coordinate R. The energy of
could be dynamically unstable. the monopole then becomes a simple polynomial in R.
A time independent spherically symmetric ansatz for the Furthermore, due to the spherical symmetry, R is a function
monopole can be chosen in the usual way as of time alone and so the original field theoretic model in
3 þ 1 dimensions reduces to a one-dimensional problem
1
KðrÞ involving RðtÞ.

a ¼ r^a hðrÞ; Aa ¼ ab r^b ; A0 ¼ 0;
er We now proceed to elucidate the existence of monopole
(5) solutions which have the thin-wall behavior described in
the previous section. Redefining the couplings appearing in
where r^ is a unit vector in spherical polar coordinates. The the potential (4) in terms of a mass scale  and expressing
energy of the monopole configuration in terms of the
in terms of the profile function hðrÞ, we have
functions h and K is
~ 2 2
Z1
02 V¼ h ðh
2 a~2 Þ2 þ 
~2 2 h2
; (10)
ðK Þ ð1
K 2 Þ2 1 2 0 2 2
EðK; hÞ ¼ 4 dr þ þ r ðh Þ
0 e2 2e2 r2 2
 1.2

þ K2 h2 þ r2 VðhÞ ; (6)
Wall
1

where derivatives with respect to r are denoted by primes.

The static monopole solution is the minimum of this func- 0.8 True Vacuum False Vacuum
tional and the ansatz functions satisfy the equations
0.6

2 2h @V
h00 þ h0
2 K2
¼ 0; (7) 0.4
r r @h
0.2

K 2
K00
ðK
1Þ
e2 h2 K ¼ 0: (8) h(r)/ η K(r)
r2 0 R
R - δ/2 R + δ/2

As r ! 1 the function h asymptotically approaches  and -0.2

r -->
is zero at r ¼ 0 from continuity requirements. On the other
hand, K approaches zero at spatial infinity so that the gauge FIG. 2 (color online). The monopole profile under the thin-
field decreases as 1=r, and K ¼ 1 at r ¼ 0. wall approximation.

025022-3
BRIJESH KUMAR, M. B. PARANJAPE, AND U. A. YAJNIK PHYSICAL REVIEW D 82, 025022 (2010)
where a tilde over a variable indicates that it is dimension-
less. The vacuum expectation value of
or h then becomes ε
~ where
, −η h0 η h -->
v
u ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
t2~ a2 a~4 ~2
3 ~2 ~
~¼
 þ : (11)
3~

−V(h) -->
3
The expression for V can be rearranged as
~a4 þ 
V ¼ ðð~ ~a2 h2 Þh2 þ Oðh6 Þ:
~2 Þ2
2~ (12)
The condition that V is approximately quadratic in h is
given by
h2 ~a4 þ 
~ ~2
 : (13)
2 ~a2
2~ FIG. 3 (color online). The scalar potential
VðhÞ which is the
Euclidean space equivalent of the potential given in (4). The
When the above condition is satisfied, @V=@h is linear in h. potential has zeroes at h ¼ h0 and h ¼ .
The equation of motion for h given in Eq. (8) can then be
written as
Eq. (4), as shown in Fig. 3. The particle must start at h ¼
2 2h 0 with a finite velocity and must reach h ¼  as r ! 1.
h00 þ h0
2
k2 h ¼ 0; (14)
r r We prove the existence of the solution that achieves h ¼
~a4 þ   at r ¼ 1 by proving that initial conditions can be chosen
where k2 ¼ ð~ ~2 Þ2 and K has been set to unity. so that the particle can undershoot or overshoot h ¼  for
Equation (14) has the form of the modified spherical Bessel r ! 1, depending on the choice of the initial velocity.
equation whose general form is Then by continuity there must exist an appropriate initial
z2 w00 þ 2zw0
½z2 þ lðl þ 1Þw ¼ 0 (15) condition for which the particle exactly achieves h ¼  at
r ! 1.
for a function wðzÞ. The primes in the above equation In the following, we will assume that K ¼ 1 is always a
denote derivatives with respect to z and Eq. (14) is obtained good approximation. Indeed, in Eq. (7) the term dependent
from (15) with l ¼ 1. on K is negligible for large r no matter the value of K,
The solution of Eq. (14) is while for small r, K ¼ 1 is a reasonable approximation. On

 the other hand, Eq. (8) for K critically depends on the value
I3=2 ðkrÞ
hðrÞ ¼ C pffiffiffiffiffi ¼ Ci1 ðkrÞ; (16) of hðrÞ Þ 0, especially for large kr. In that sense, the
kr function hðrÞ does not depend strongly on KðrÞ, whereas
where IJ is the modified Bessel function of the first kind of hðrÞ drives the behavior of KðrÞ.
order J, in is the modified spherical Bessel function of the
first kind of order n, and C is an arbitrary constant. The A. Overshoot
function i1 ðkrÞ  ekr =ðkrÞ for kr  1 and is linear in kr The existence of the overshoot can be proven by taking a
for small kr  1. If we choose C ¼ e
k
with arbitrarily sufficiently small value of C. As explained earlier, C can be
large k
, we see that we can keep Eq. (13) satisfied and chosen small enough so that Eq. (13) is valid even for large
hence stay with the linear equation for hðrÞ for arbitrarily kr; hence, the equation remains linear. If kr is large
large kr. enough, the friction term ð2=rÞh0 and the term ð2=r2 Þh in
The existence of the particular solution with hðrÞ ¼  at the equation of motion can be neglected in any further
r ¼ 1 can be proven using an argument similar to evolution and the evolution can be thought of as conserva-
Coleman’s, where he proved, in a somewhat different tive. Thus with such a choice of C, h increases to h~ < h0 at
context, the existence of a thin-wall instanton, [15]. We a large value of kr according to the linearized equation (h0
can reinterpret the equation for the monopole profile, is the zero crossing point of the potential; see Fig. 3). The
Eq. (7), as describing the motion of a particle whose motion from then onwards is frictionless. The particle has
position is denoted by hðrÞ where r is now interpreted as ~ thus its energy is still positive
an energy E > 0 at h ¼ h;
a time coordinate. The particle moves in the presence of
when it reaches h ¼ . As a result, it overshoots to h > .
friction with a time dependent Stokes coefficient given by
the second term in Eq. (7) and a time dependent force given
by the third term in Eq. (7) (setting K ¼ 1), both of which B. Undershoot
are singular at r ¼ 0. The particle also moves in the To prove the existence of the undershoot, we start with
potential
VðhÞ, obtained by inverting the potential the full equation for hðrÞ:

025022-4
FATE OF THE FALSE MONOPOLES: INDUCED VACUUM DECAY PHYSICAL REVIEW D 82, 025022 (2010)
2 2 @V 1. Overshoot
h00 þ h0
2 h
¼0 (17)
r r @h For the case of the overshoot, we use the expression
Eq. (19) which gives
which after multiplying both sides by h0 can be rewritten as
Z rf
0


0  E ¼
2 drh0
h
d 1 02 h h : (22)
ðh Þ
VðhÞ ¼
2h0
2 ; (18) r0 r
dr 2 r r
Assuming that h0 ðrÞ is positive, we will find an estimate for

0 h0 ðrÞ < v. Then
h 
¼
2h0 : (19) 

 Z rf
h0  


r jEj < 2v  dr 
 (23)

 r  
r0
The quantity on the left-hand side of Eq. (18) can be



thought of as the time derivative of the energy E. In the 

 hðrf Þ hðr0 Þ 



linearized regime, it is easy to show that the right-hand side ¼ 2v

 (24)
is strictly negative for all r. It starts with a value of zero at
 rf r0  
r ¼ 0 and decreases essentially exponentially for large kr. 



  hðr0 Þ 
 
We can choose C, which amounts to choosing the initial < 2v

 r

; (25)
velocity so that h evolves according to the linearized  f r0 
equation until kr can be taken to be large. However, in
contrast to the case of the undershoot, we now require that where we replaced hðrf Þ with  since that is its largest
E becomes negative. This means that the value of C is possible value. As long as v is well behaved, as r0 ! 1,
taken larger than in the case of the overshoot. E is made up rf > r0 , thus the first term vanishes, while the second term
of two terms, the kinetic term which is positive semidefin- can be made small by choosing the value of C to be
ite and the potential term which becomes negative for h > arbitrarily small. Thus we see that E ! 0 and therefore
h0 . We impose conditions on the parameters so that E the change in the energy is arbitrarily small. Thus we
becomes negative and consequently h > h0 within the necessarily obtain an overshoot since at r ¼ r such that
linearized regime. Now if kr is large enough, as before, hðr Þ ¼ , VðÞ ¼ 0; hence the particle has a positive
the subsequent evolution will be conservative, and since kinetic energy giving an overshoot.
the total energy is negative, the subsequent evolution will To get the value of v, we use Eq. (18):
never be able to overcome the hill at h ¼  and the particle

0 
d 1 02 h h
will undershoot. ðh Þ
VðhÞ ¼
2h0
2 ; (26)
dr 2 r r
C. Technical details
hh0 ðh2 Þ0
To make the previous arguments more precise and rig- <2 < 2 : (27)
r2 r0
orous, we note that when the condition Eq. (13) is satisfied,
the linear regime is valid and VðhÞ is approximately qua- Integrating both sides from r0 to rf yields
dratic in h; i.e.,
VðhÞ  
ð1=2Þk2 h2 and the equation

1
of motion for h is approximately ðh ðrf ÞÞ < 2 2 ðh2 ðrf Þ
h2 ðr0 ÞÞ
0 2
(28)
r0


0 
d 1 02 1 h h
ðh Þ þ 
k2 h2 ¼
2h0
2 : (20) 
dr 2 2 r r 1 0 2
þ Vðhðrf ÞÞ
Vðhðr0 ÞÞ þ ðh ðr0 ÞÞ : (29)
2
Using the properties of i1 ðkrÞ we can compute E in the
linear regime; we find for large kr Thus v2 is given by


1 2
k2 C2 e2kr v2 ¼ 2 ð
h2 ðr0 ÞÞ (30)
E
(21) r20
4ðkrÞ3

which can be evidently taken to be positive or negative by 1
þ supjVðhðrf ÞÞ
Vðhðr0 ÞÞj þ ðh0 ðr0 ÞÞ2 (31)
simply choosing the value of C. Then in the subsequent 2
evolution, where we can no longer rely on the linear
which is a bounded function of r0 .
evolution, the right-hand side has two competing terms:
the friction term, which only reduces the energy, and the
time dependent force term, which tries to increase it. The 2. Undershoot
change in the energy for evolution between r0 and rf is To prove the undershoot we use the expression Eq. (18)
given by the integral of the right-hand side. which gives

025022-5
BRIJESH KUMAR, M. B. PARANJAPE, AND U. A. YAJNIK PHYSICAL REVIEW D 82, 025022 (2010)
Z rf h02 Z rf h0 h 2
E ¼
2 dr þ2 dr 2 : (32) < h2 ðr0 Þ < 2 : (42)
r0 r r0 r kr0
Integrating the second term by parts we obtain Using hðr0 Þ  Cekr0 =2kr0 we can choose
Z rf h0 h Z rf
h2 0 Z rf
2h2  2kr0
2 dr 2 ¼ dr 2 þ dr 3 (33) C¼ (43)
r0 r r0 r r0 r ekr0 r1=4
0

2 
 which gives
h  


rf
2 1 


rf
< 2  
  ;
2 
(34) 2 2
r  r 
r0 r0 < pffiffiffiffiffiffiffi < 2 : (44)
kr0 kr0
where we obtain the inequality using the fact that we are
only interested in the region h  . It is obvious that for large enough kr0 this is easily sat-
We now prove that this contribution to the energy cannot isfied. Thus we have established the existence of a choice
be sufficient to push h to h > . We take r0 to be the value of C or initial velocity which contradicts the existence of
of r as described after Eq. (19), where the energy becomes r .
negative within the linearized regime with kr0  1. We
now assume there exists a value rf  r for which hðr Þ ¼
IV. COLLECTIVE COORDINATE AND THE
. Then INSTANTONS
Z r h02
h2  r

 r

E <
2 þ 2 


 2 1   The potential Vð
Þ given in (4) can be normalized so
dr 
  (35)
r0 r r  r0 r 
2  r0 that the energy density of the metastable vacuum is vanish-
ing whereas the energy density of the true vacuum is
.


2 h2 ðr0 Þ 2 1 1 By making use of the thin-wall approximation, the expres-
< 2
2
 2
2 (36) sion for the total energy in the static case given in (6) can
r r0 r r0
be expressed as
Z R
ð =2Þ Z1
2
h2 ðr0 Þ 1
¼ (37) E ¼ 4 drr2 VðhÞ þ dr 2 2
r20 0 Rþð =2Þ 2e r
Z Rþð =2Þ
ðK0 Þ2 ð1
K 2 Þ2 1
which is an upper bound to the energy that can be added to þ dr þ þ r2 ðh0 Þ2
the particle. But now it is easy to see that this additional R
ð =2Þ e2 2e2 r2 2

energy is insufficient to push h to h > , for kr0 large
þ K 2 h2 þ r2 VðhÞ : (45)
enough. Indeed the energy of the particle at r ¼ r0 is
obtained, via the linear regime, by Eq. (21):
In the above expression, we have made use of the fact that
k2 C2 e2kr h2 ðr0 Þ VðhÞ is zero for r > R þ 2 , K ¼ 1 for r < R
2 , K ¼ 0
E
!
k : (38)
4ðkrÞ 3 r0 for r > R þ 2 , and both the derivative terms and the term
K2 h2 are nonzero only when R
2 < r < R þ 2 . Since
This expression is negative. Furthermore, if kr0 is large
is small, the first integral on the right-hand side of (45)
enough, we will see that E cannot provide enough energy
gives
R3 where  ¼ 4=3 because VðhÞ ¼
 in the
to increase E to zero, giving a contradiction to the exis-
domain of integration. The second integral gives C=R
tence of r . To see this, we would require jEj > E, i.e.,
where C ¼ 2=e2 . The third integral is due to the energy
h2 ðr0 Þ 2
h2 ðr0 Þ of the wall and can be written as 4R2 where  is the
k
> : (39) surface energy density of the wall given by
r0 r20

02
1 Z Rþð =2Þ ðK Þ ð1
K2 Þ2 1 2 0 2
The linear approximation assumes hðr0 Þ  ; hence we ¼ 2 dr þ þ r ðh Þ
get R R
ð =2Þ e2 2e2 r2 2

kh2 ðr0 Þ 2 þ K2 h2 þ r2 VðhÞ : (46)

2 > (40)
r0 r0
We can thus write the total energy of the monopole as
reorganizing the terms, which for small enough  simply
C
implies EðRÞ ¼
R3 þ 4R2 þ : (47)
R
2 2
h ðr0 Þkr0 >  : (41)
This function is plotted in Fig. 4. There is a minimum at
Thus we get the inequality sandwich R ¼ R1 and this corresponds to the classically stable

025022-6
FATE OF THE FALSE MONOPOLES: INDUCED VACUUM DECAY PHYSICAL REVIEW D 82, 025022 (2010)

2
1 _a _a 1 dK _ 2
A A ¼ 2 2 R: (52)
4 2e r dr
The Lagrangian can then be expressed as
Z 1

dh2
 
1 dK 2 _ 2
R_ 2 þ 2
ENERGY -->

L ¼ 2 r2 R dr
EðRÞ:
0 dr e dr
Instanton Tunneling
(53)
From (8), for large r, the equation of motion of h can be
written as
R1 R2
@VðhÞ
h00
¼ 0: (54)
@h
RADIUS OF BUBBLE -->
Multiplying both sides by h0 and integrating by parts with
FIG. 4 (color online). The function EðRÞ plotted versus bubble respect to r, one obtains
radius. The classically stable monopole solution has R ¼ R1 . pffiffiffiffiffiffiffiffiffiffiffiffiffi
This solution can tunnel quantum mechanically to a configura- h0 ¼ 2VðhÞ: (55)
tion with R ¼ R2 and then expand classically.
Furthermore, since dh=dr is nonvanishing only in the thin-
wall, the value of r in the first integral in (53) can be
monopole solution. This solution has a bubble of true replaced by R and we have
vacuum in its core and the radius R1 of this bubble is Z1
2 Z 1
dhpffiffiffiffiffiffiffiffiffiffiffiffiffi
obtained by solving dE=dR ¼ 0. However, this monopole 2 dh _ 2 2 _2
drr R ¼R R dr 2VðhÞ
configuration can tunnel quantum mechanically through 0 dr 0 dr
the finite barrier into a configuration with R ¼ R2 where ¼ R2 R_ 2 S1 ; (56)
EðR1 Þ ¼ EðR2 Þ. Once this occurs, the monopole can con-
tinue to lose energy through an expansion of the core since where
the barrier which was present at R1 is no longer able to Z pffiffiffiffiffiffiffiffiffiffiffiffiffi
prevent this. S1 ¼ dh 2VðhÞ: (57)
We now proceed to determine the action of the instanton 0
describing the tunneling from R ¼ R1 to R ¼ R2 . In the Defining
thin-wall approximation, the functions h and K can be
2
written as 1 Z1 dK
S2 ¼ 2
dr ; (58)
e 0 dr
h ¼ hðr
RÞ; K ¼ Kðr
RÞ; (48)
the Lagrangian (53) becomes
and the exact forms of the functions h and K will not be
required in the ensuing analysis. The only requirement is L ¼ 2R_ 2 ðS1 R2 þ S2 Þ
EðRÞ (59)
that both h and K change exponentially when their argu- and the action can be written as
ment (r
R) is small. An example of a function with this
type of behavior is the hyperbolic tangent function. The Z1
S¼ dtð2R_ 2 ðS1 R2 þ S2 Þ
EðRÞÞ: (60)
time derivative of
can be written as
1

dh _ In Euclidean space, the expression for the action becomes

_ a ¼ r^a R: (49)
dR Z1
SE ¼ d ð2R_ 2 ðS1 R2 þ S2 Þ þ EðRÞÞ; (61)
From (48), since ðdh=dRÞ2 ¼ ðdh=drÞ2 , we have
1


 where ¼ it is the Euclidean time and R_ is the derivative

1 _ a _ a 1 dh 2 1 dh 2 _ 2


¼ ; R_ 2 ¼ R : (50) with respect to . The instanton solution Rð Þ which we are
2 2 dR 2 dr
seeking obeys the boundary conditions R ¼ R1 for ¼
Similarly, 1, R ¼ R2 for ¼ 0, and dR=d ¼ 0 for ¼ 0. It can
be obtained by solving the equations of motion derived



1 dK _ from (61). However, the exact form for Rð Þ will not be of
A_ a ¼ ab r^b R (51) interest here since the decay rate of the monopole is
er dR
determined ultimately from SE [15]. The calculation of
and SE will be the subject of the next section.

025022-7
BRIJESH KUMAR, M. B. PARANJAPE, AND U. A. YAJNIK PHYSICAL REVIEW D 82, 025022 (2010)
V. BOUNCE ACTION 0 neglecting the constant and linear term in R since R2 is
large, obtaining R2  4= ¼ 3=.
In this section, we will derive an expression for bounce
Factoring J, we have
action SE for the monopole tunneling and compare it with
the bounce action for the tunneling of the false vacuum to pffiffiffiffi Z R2 pffiffiffiffiqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
the true vacuum as discussed in [15] with no monopoles SE ¼ A  dR R
ðR
R1 Þ2 ðR
R2 ÞðR
R3 Þ
R1
present. From (61), the equation of motion for R can be pffiffiffiffi Z R2 pffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
written ¼A  dR RðR
R1 Þ
ðR
R2 ÞðR
R3 Þ
R1
1 @E


ðR S1 þ S2 ÞR€ þ S1 RR_ 2

2
¼ 0: (62) pffiffiffiffi 2 R1 5=2 R1 R3
4 @R ¼ A R7=2 2 1
I ;
105 R2 R2 R2
Multiplying both sides by R,_ the equation of motion as-


pffiffiffiffiffiffiffiffiffiffiffiffiffiffip ffiffiffiffiffiffiffiffiffiffiffiffiffiffi 7=2 2 R1 5=2 R1 R3
sumes the form ¼ 32S1 4=3R2 1
I ; :
  105 R2 R2 R2
d 1 EðRÞ
ðS2 þ R2 S1 ÞR_ 2
¼ 0: (63) (69)
dt 2 4
Here I is a dimensionless function of R1 =R2 and R3 =R2
The term in the square brackets is a constant of motion and which is finite everywhere in the domain ½R1 ; R2  and is
can be taken to be zero with loss of generality. Setting this obtained from the integral defined in Eq. (69) removing the
constant to zero gives
factor of ð1
ðR1 =R2 ÞÞð5=2Þ and R27=2 and some numerical
EðRÞ ¼ 2ðS2 þ S1 R2 ÞR_ 2 : (64) factors. It is expressible in terms of elliptic integrals and its
explicit expression is not illuminating. As S1 has dimen-
Substituting this in (61), we have sions of 3 and  has dimensions of 4 , the expression is
Z1 dimensionless, as expected. Substituting the value of R2 in
SE ¼ d 4ðS2 þ S1 R2 ÞR_ 2 : (65) SE ,

1



Solving for R_ from (64) and using this in the above 144 pffiffiffiffiffiffiffiffi 7=2 R 5=2 R1 R3
SE ¼ 2S1 3 1
1 I ; : (70)
equation yields 35  R2 R2 R2
Z1

dR ~ in the potential (10) can
SE ¼ d 4ðS2 þ S1 R2 ÞR_ For small , the term containing 

1 d be neglected. Using Eq. (57) and the fact that  ¼ a~
pffiffiffiffiffiffiffiffiffi Z R2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi when ~ ¼ 0,
¼ 32 dR ðS2 þ S1 R2 ÞEðRÞ: (66) pffiffiffiffiffiffi
2~ Z a~
R1

Using the expression for EðRÞ given in (47) and neglecting S1 ¼ dhðhðh2
2 a~2 ÞÞ (71)
 0
S2 in comparison to S1 R2 , the Euclidean action of the
bounce solution can be written sffiffiffiffi
Z R2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ~ 4 3
¼ a~  : (72)
SE ¼ A dR
ðR5
4R4
CR þ E0 R2 Þ; (67) 8
R1
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi The value of  can be obtained from Eq. (46) by noting that
where A ¼ 32S1 . In deriving the above expression, the
the terms multiplying r2 are large compared to the terms
constant E0 ¼ EðR1 Þ was subtracted from the expression
independent of r and the term multiplying 1=r2 . Since is
for EðRÞ in (47) so that the bounce has a finite action.
small, we can write r ¼ R and Eq. (46) becomes
Pulling out a factor of R from the square root in the
integrand, we have Z Rþð =2Þ
1 
¼ 0 2
dr ðh Þ þ VðhÞ : (73)
Z R2 pffiffiffiffipffiffiffiffiffiffiffi R
ð =2Þ 2
SE ¼ A dR R
J ; (68)
R1 Substituting for h0 from Eq. (55),  becomes
4 3
where J ¼ R
4R
C þ E0 R. The function J has Z Rþð =2Þ
a double root at R ¼ R1 , a positive root at R ¼ R2 , and a ¼ drðh0 Þ2 (74)
R
ð =2Þ
negative root at R ¼ R3 . Since we are working with  small
and  ¼ 4=3, we can neglect the term containing  Z
while solving dE=dR ¼ 0 and obtain R1  ð4e2 Þ
1=3 . To ¼ dhðh0 Þ (75)
find R3 we also neglect the term containing , and sub- 0

stituting for E0 in terms of the solution for R1 , we get a

Z pffiffiffiffiffiffiffiffiffiffiffiffiffi
cubic equation for R3 , which can be exactly factored, ¼ dh 2VðhÞ (76)
giving R3 ¼
2R1 . Finally, to solve for R2 , we solve J ¼ 0

025022-8
FATE OF THE FALSE MONOPOLES: INDUCED VACUUM DECAY PHYSICAL REVIEW D 82, 025022 (2010)
¼ S1 : (77) bles will eventually convert the entire system to the true
vacuum without the need for quantum tunneling. Such a
Using (77) and (72) in (70) yields phenomenon may be referred to as a rollover transition
pffiffiffi

 [20] characterized by the relevant critical value.
144 2 S41 R1 5=2 R1 R3
SE ¼ 1
I ; (78)
35 3 R2 R2 R2
VI. MONOPOLE DECAY IN A SUPERSYMMETRIC
pffiffiffi

 SUð5Þ GRAND UNIFIED THEORY MODEL
9 2 ~2 16 12 R1 5=2 R1 R3
¼  a~ 1
I ; (79)
140 3 R2 R2 R2 The results of this work have direct relevance to a super-
symmetric SUð5Þ model studied in [25] in which super-
as the final value of the bounce action. From the values of symmetry symmetry breaking is sought directly through
R1 and R2 , we have O’Raifeartaigh type breaking. The Higgs sector, which
R1 1 1  contains two adjoint scalar superfields 1 and 2 and the
¼ 2=3 (80) superpotential, including leading nonrenormalizable
R2 e ð4Þ 3
1=3
terms, is of the form

1=3 

1 16   
W ¼ Tr 2 1 þ 21 þ 1 31 þ 2 Trð21 Þ1
¼ ; (81)
~
ðeÞ 2=3 27
a~ 4
16=3 M M


 21
where the value of  has been expressed in terms of the ¼ 1 2 
p ffiffiffiffiffi
ffi 1 þ ð71 þ 302 Þ ; (85)
couplings appearing in the potential using Eqs. (77) and 30 30M
(72). From the expression given in (79), it is evident that where 1 and 2 are selected components of 1 and 2 ,
the bounce action SE is zero when R1 ¼ R2 as expected. respectively, relevant to the symmetry breaking. Two mass
With  small, R1 =R2 is small, but it is interesting to note scales appear in the superpotential,  and M, the latter
that variations in the couplings can reduce the bounce being a larger mass scale whose inverse powers determine
action. For example, a reduction in the Uð1Þ gauge cou- the magnitudes of the coefficients of the nonrenormaliz-
pling e has the effect of increasing the monopole mass and able terms. The scalar potential derived from this super-
of reducing the bounce action. potential can be written as
We now compare our answer with the well known

2 7 3  3 2
formula of [15] relevant to homogeneous nucleation, i.e., V ¼ 1
pffiffiffiffiffi1ffi þ 1 1 þ 2 1
tunneling of the translation invariant false vacuum to the 30 30M M



true vacuum. Denoting this bounce to be B0 , 21 ð71 þ 302 Þ 2 2
þ 2 
pffiffiffiffiffiffi þ 1 : (86)
272 S41 30 10M
B0 ¼ (82)
2 3 In [22], monopole solutions were shown to exist in this
model and the classical instability of the vacuum structure
272 ~2 16 12 of this theory in the presence of such monopoles was
¼  a~ : (83)
128 3 discussed.
Thin-walled monopoles can be obtained in this model
Comparing this expression with our bounce B  SE for the
under the condition
monopole assisted tunneling given in (79), we see that pffiffiffiffiffiffi
pffiffiffi

 1 30
32 2 R 5=2 R1 R3  (87)
B¼ B0 1
1 I ; : (84)  2
105 R2 R2 R2
which is equivalent to the condition in Eq. (13), and hence
We see that unlike the homogeneous case, the bounce can the results of this paper could be applied directly there. In
parametrically become indefinitely small and vanish in the [22] the region of parameter space studied did not coincide
limit R1 ! R2 . The interpretation of this limit is that the with this condition, and thus the monopoles were not thin
very presence of a monopole in this parameter regime walled. The monopoles were classically unstable when
implies the unviability of a state asymptotically approach-   M4 was increased beyond a critical value. We can
ing the vacuum deduced by a naive use of the effective recover this behavior from Eq. (79) as  is increased;
potential. If the parameters in the effective potential ex- however, it is important to note that our approximation in
plicitly depend on external variables such as temperature, it this paper becomes invalid for large enough .
may happen that the limit R1 ! R2 is reached at a critical
value of this external parameter. In this case, as the external
VII. DISCUSSIONS AND CONCLUSIONS
parameter gets tuned to this critical value, the monopoles
will become sites where the true vacuum is nucleated We have calculated the decay rate for so-called false
without any delay and the indefinite growth of such bub- monopoles in a simple model with a hierarchical structure

025022-9
BRIJESH KUMAR, M. B. PARANJAPE, AND U. A. YAJNIK PHYSICAL REVIEW D 82, 025022 (2010)



of symmetry breaking. The toy model that we use has a R1 5=2 R1 R3
breaking of SUð2Þ to Uð1Þ which is the false vacuum, F ðR1 ; R2 ; R3 Þ ¼ R7=2
2 1
I ; ; (89)
R2 R2 R2
which in principle happens at a higher energy scale, and
then a true vacuum which has no symmetry breaking. The
symmetry broken false vacuum admits magnetic mono- where  contains the determinantal and zero mode factors,
poles. The false vacuum can decay via the usual creation of and I is defined in Eq. (69). In the limit that R1 ! R2 the
true vacuum bubbles [15]; however, we find that this decay tunneling rate is unsuppressed while the homogeneous
can be dramatically enhanced in the presence of magnetic tunneling rate for the nucleation of true vacuum bubbles
monopoles. Even though the false vacuum is classically as found by Coleman [15] still remains suppressed. Hence
stable, the magnetic monopoles can be unstable. At the in this limit, the classical false vacuum is classically stable,
point of instability, the monopoles are said to dissociate. but subject to quantum instability through the nucleation of
This corresponds to an evolution where the core of the true vacuum bubbles, but the rate for such a decay can be
monopole, which contains the true vacuum, dilates indef- quite small. However, the existence of magnetic monopole
initely [17,18,24]. However, before the monopoles become defects renders the false vacuum unstable, and in the limit
classically unstable, they can be rendered unstable from of large monopole mass, the decay rate is unsuppressed.
quantum tunneling. We have computed the corresponding
rate and find that as we approach the regime of classical ACKNOWLEDGMENTS
instability, the exponential suppression vanishes. The tun-
neling amplitude behaves as We thank NSERC, Canada for financial support. The
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi visit of B. K. was made possible by a grant from CBIE,

   Canada. The research of U. A. Y. is partly supported by a
  16 2S1 2 
 exp F ðR1 ; R2 ; R3 Þ (88) grant from DST, India. The authors would like to thank
V 2 105 3
R. MacKenzie and P. Ramadevi for useful comments re-
with garding this work.

[1] G. ’t Hooft, Nucl. Phys. B79, 276 (1974). [13] S. Kachru, R. Kallosh, A. D. Linde, and S. P. Trivedi, Phys.
[2] A. M. Polyakov, JETP Lett. 20, 194 (1974). Rev. D 68, 046005 (2003).
[3] A. M. Polyakov, Nucl. Phys. B120, 429 (1977). [14] I. Y. Kobzarev, L. B. Okun, and M. B. Voloshin, Sov. J.
[4] S. R. Coleman, Subnuclear series 13, 297 (1977). Nucl. Phys. 20, 644 (1975).
[5] R. Rajaraman, Solitons and Instantons. An Introduction to [15] S. R. Coleman, Phys. Rev. D 15, 2929 (1977).
Solitons and Instantons in Quantum Field Theory (North- [16] S. R. Coleman and F. De Luccia, Phys. Rev. D 21, 3305
Holland, Amsterdam, 1982). (1980).
[6] J. Preskill and A. Vilenkin, Phys. Rev. D 47, 2324 (1993). [17] P. J. Steinhardt, Nucl. Phys. B190, 583 (1981).
[7] T. H. R. Skyrme, Nucl. Phys. 31, 556 (1962). [18] Y. Hosotani, Phys. Rev. D 27, 789 (1983).
[8] T. Gisiger and M. B. Paranjape, Phys. Rep. 306, 109 [19] U. A. Yajnik, Phys. Rev. D 34, 1237 (1986).
(1998). [20] U. A. Yajnik and T. Padmanabhan, Phys. Rev. D 35, 3100
[9] See [10] for discussion of supersymmetric field theories. (1987).
[10] J. Terning, in Particle Physics and Cosmology: The Quest [21] B. Kumar and U. A. Yajnik, Phys. Rev. D 79, 065001
For Physics Beyond The Standard Model(s), edited by (2009).
H. E. Haber and A. E. Nelson (TASI, Boulder, CO, 2002). [22] B. Kumar and U. Yajnik, Nucl. Phys. B831, 162 (2010).
[11] M. Dine and A. E. Nelson, Phys. Rev. D 48, 1277 [23] S. R. Coleman, Subnuclear series 15, 805 (1979).
(1993). [24] P. J. Steinhardt, Phys. Rev. D 24, 842 (1981).
[12] K. A. Intriligator, N. Seiberg, and D. Shih, J. High Energy [25] B. Bajc and A. Melfo, J. High Energy Phys. 04 (2008) 062.
Phys. 04 (2006) 021.

025022-10