The Extremal Higher Dimensional Constructions of

Fundamental Brane–Antibrane Systems

Boris Stoyanov

DARK MODULI INSTITUTE, Membrane Theory Research Department,

18 King William Street, London, EC4N 7BP, United Kingdom
E-mail: [email protected]

BRANE HEPLAB, Theoretical High Energy Physics Department,

East Road, Cambridge, CB1 1BH, United Kingdom
E-mail: [email protected]

SUGRA INSTITUTE, 125 Cambridge Park Drive, Suite 301,

Cambridge, 02140, Massachusetts, United States
E-mail: [email protected]

Abstract
We calculate the extremal higher dimensional effective actions of fundamental brane-antibrane
systems elegantly presented in the theoretical framework of advanced membrane theory con-
structions. Detailed study of brane-antibrane systems reveals when brane separation is smaller
than the superstring length scale, spectrum of this system has different tachyonic modes and
interaction regimes in the moduli superspace. The higher dimensional effective actions should
then include these modes because they are the most important ones which rule the extremal
dynamics of the fundamental brane systems. In this regard, it has been shown that an effec-
tive action of Born-Infeld type proposed in the current literature can capture many properties
of the decay of non-BPS Dp-branes in superstring and membrane theory. The effective actions
of brane-antibrane systems in Types IIA and IIB superstring theories should be given by some
extension of the DBI action and the WZ terms which include the tachyon field configurations.
The DBI part may be given by the projection of the effective action of two non-BPS Dp-branes
in Type IIB theory. We are interested in this paper in the appearance of tachyon, gauge field and
the RR field in these extremal higher dimensional actions. Using the consistency of the present
constructions, we have also found the first higher derivative corrections to the exceptional part
of the extremal effective actions with brane-antibrane systems.
1 Introduction
In the presented doctrina dominum article, we will focus our efforts and concentrate on building
a perfect complete picture of the extremal higher-dimensional fundamental interactions of brane-
antibrane systems with the exclusive inclusion of superstring theory principles plus magical set
of supergravity consistency conditions with the dream aiming to place an advanced degree in
the formation of a realistic membrane theory. We examine the low energy effective field theory
following from the tachyon effective action in the background of static brane-antibrane systems,
neglecting the gravitational and Ramond-Ramond fields. In the limit that the thickness of the
branes tends to zero, and the tachyon field between them approaches the superstring vacua,
we should expect interactions between the branes to vanish and the resulting spectrum should
represent the effective field theory around isolated branes. It has recently been realized that
brane-antibrane annihilation may result in defect formation, due to the dynamics of the tachyon
field. Studies of this possibility have generally ignored the interaction of the brane fields with
fields in the bulk, recently it has been argued that interactions with bulk fields suppress or even
eliminate defect formation. The usual approach to defect formation during tachyon condensation
is to take into account only the fields that live in the worldvolume of the decaying non-BPS brane
or brane-anti-brane pair. This approach seems to be motivated by the fact that the resulting
lower-dimensional branes formed during the decay are localized inside the worldvolume of the
parent brane. However, the final state defects are themselves D-branes, and therefore couple to
bulk RR-fields. One should include the effects of these fields in the defect formation process. The
open string tachyon condensation on non-BPS brane systems has attracted much interest recently.
One framework of analysis is level truncation of the open string field theory (SFT) which lead to
very good numerical agreements with expected values of vacuum energy and lower-dimensional
D-branes tensions. Another exceptional theoretical framework is the boundary SFT (BSFT). It
was argued that while in the SFT approach an infinite number of massive fields are involved in the
condensation process, in the BSFT one can restrict to the tachyon field and study some aspects
of the condensation, such as the tensions of the lower-dimensional D-branes, exactly. Recently, it
has been realized that the spectrum in some vacua of type II string theories contains not only BPS
D-branes but also unstable non-BPS D-branes. And the in- stability of these non-BPS D-branes
has been interpreted in the stringy context in terms of the tachyonic mode arising in the spectrum
of open strings ending on the non-BPS D- branes. Indeed, there could be several ways in which
one can argue why the tachyonic mode should develop in the spectrum of open strings ending
on unstable D-branes such as brane-antibrane systems or non-BPS D-branes with unidentified
worldvolume dimensions in IIA/IIB superstring theories. In this note we will use the notion of
superconnections, which when considering the branes-antibranes system and non-BPS Dp-branes
appears naturally via the Chan-Paton factors. We will make the assumption that the effective
action of tachyon and gauge fields for the brane-antibrane system and non-BPS Dp-branes can be
written in a Quillen-like framework in terms of the supercurvature. The tachyon potential that
arises in this framework is exponential in the tachyon field. We will propose a form of the effective
action and use it to study the process of tachyon condensation. Kink solutions that we will find,
with infinite constant value of the gauge field strength, reproduce the exact tensions of the lower-
dimensional D-branes at the minimum of the tachyon potential. The effective action is different
from the BSFT proposal. It can be related by field redefinitions, in some cases, to the extremal
effective action proposed in the article. In superstring theory a pair of parallel brane-antibrane
system constitutes an unstable object. To study the dynamics of unstable D-branes, the BSFT
is a useful tool and it has provided a good understanding of tachyon condensation at the classical
level. It describes the off-shell dynamics of open strings in a fixed on-shell background of closed
strings in which an open string field configuration corresponds to a boundary term in the world-
sheet action of the string. Therefore, specifying a boundary term means giving the background
values of the various modes of the open string. It is based on the Batalin-Vilkovisky formalism

2
whose master equation provides the effective action of the theory. The Tachyon-Dirac-Born-
Infeld (TDBI) action captures some aspects of the dynamics of non-BPS D-branes in type II
string theory. We show that it can also be used to study the classical interactions of BPS branes
and antibranes. Our analysis sheds light on real time tachyon condensation, on the proposal that
the tachyon field can be thought of as an extra spatial dimension whose role is similar to the
radial direction in holography, and on A. Sen’s open string completeness conjecture. However, it
is easy to see that the numerical coefficients of the higher derivative terms do not match those
of the BSFT action. Indeed, in the current literature one studies the tachyon condensation with
only the tachyon field excited and one gets the precise tension of the lower-dimensional D(p-1)-
brane. In our variables, we needed a nonzero configuration of the gauge field strength in order
to derive the precise tension of D(p-1)-brane from the kink solution. Upon addition of the gauge
fields in the BSFT formalism there is still a difference between the actions. We try to analyze
the mechanism of defect formation when bulk fields are involved. We first construct the extremal
model which captures the essential features of the full 10-dimensional model that are responsible
for the formation of the topological defects. Using this special model we study the classical
evolution of the brane and bulk fields, so the deSitter quantum fluctuations are responsible only
for seeding perturbations which grow during the subsequent evolution. The final configuration
of the bulk field gradients is a result of this classical evolution. The energy of the system comes
from tension of the brane-antibrane pair, which can be modeled as the potential energy of the
tachyon field sitting at the top of the potential. In our extremal model this corresponds to the
potential energy of a complex scalar living in the worldvolume of the fundamental brane-antibrane
system. We will present the supergravity Lagrangian for the fields which live both inside the
worldvolume of the brane-antibrane system and the bulk, and we will construct a extremal model
with the same physical features. We examine the low energy effective field theory following from
the tachyon effective action in the background of static brane-antibrane systems, neglecting the
gravitational and Ramond-Ramond fields. Moreover, recently the interest in the nature of this
tachyon potential has been multipled by the idea of unstable brane decay via the open string
tachyon condensation on these unstable Dp- branes. The conjecture according to which the
tachyon potential has a minimum and the negative energy density contribution from the tachyon
potential at the minimum should exactly cancel the sum of the tensions of the brane-antibrane
system. Thus the endpoint of the unstable brane-antibrane annihilation should be a closed
superstring vacua without any D-brane. In the present work, we shall attempt to read off the
supergravity analogue of tachyon potential from the semi-classical supergravity description for
brane-antibrane interaction. To this end, we begin with the brief review of the conjecture for
tachyon condensation in the brane-antibrane annihilation process which can be stated as follows.
We start this study here with the examination of the equations of motion on parallel branes which
are possible for fundamental test of brane-antibrane systems. These are among the simplest to use
in applications, partly because of the triviality of the various consistency conditions which would
otherwise arise when trying to construct brane configurations inside compact extra dimensions.
The dynamic process of unstable D-branes decaying into stable ones with one dimension lower
can be described by a tachyon field with a Dirac-Born-Infeld effective action. We investigate
the fluctuation modes of the tachyon field around a two-parameter family of static solutions
representing an array of brane-antibrane systems. We extended the effective action of a Dp-
brane via its tangential and transverse dynamics. In fact, there are various extensions for the
brane effective action: derivative corrections, tachyonic extension and curvature corrections.
Accordingly, for a given setup of a brane-antibrane systems one may combine some of these
modifications to construct a suitable higher-dimensional effective interaction. For a dynamical
brane-antibrane system with the inclusion of tachyon fields in a curved supergravity background
one should add an appropriate tachyon potential and the supercurvature improvements to the
specific action. The presented research work suggests further investigations in many different
directions in establishing a unified framework for fundamental membrane theory.

3
2 The Equations Relevant to Motion of Parallel Branes
We start by setting up and solving the equations which are relevant to the motion of a probe
antibrane moving within the background of N parallel branes. To do so we first examine the
fields which the N source branes set up, and then examine the equations of motion which they
imply for a test brane or antibrane. We show how these equations may be solved for the motion
of the test-brane centre of mass when all other brane modes are frozen. Let us first consider the
gravitational, dilaton and electromagnetic fields which are set up by a set of N parallel p-branes
in D-dimensional spacetime. We take the action for the fields to be given in Einstein Frame by
Z
D √  1 MN   1 αφ M1 ...Mn

Ss = − d x −g g RM N + ∂M φ ∂N φ + e FM1 ...Mn F , (2.1)
2 2 n!
where the n-form field strength is related to its (n − 1)-form gauge potential in the usual way
F[n] = dA[n−1] . Here n is related to the spacetime dimension d of the N p-branes by d = p + 1 =
n − 1. We denote the d coordinates parallel to the branes by xµ and the D − d transverse
coordinates by y m .
The constant α depends on which kind of brane is being considered. For instance, if A[n−1]
were to arise from a string-frame action
1 Z D q
SSF = − d x −gb eαs φ FMb1 ...Mbn F M1 ...Mn , (2.2)
b b
2 n!
with the string-frame metric given by gbM N = eλφ gM N for λ = 4/(D − 2), then α = αs + 2 − λ(n −
1). Two cases of particular interest are: (i) NS-NS fields (like the fields rising in the gravity
supermultiplet in various dimensions), for which αs = −2 and so α = αN S = −4(n − 1)/(D − 2);
(ii) R-R fields, for which αs = 0 and so α = αR = 2(D − 2n)/(D − 2). In the special case
D = 10 we have λ = 12 and so these two cases become αN S = − 12(n − 1) = − 12(p + 1) or
αR = 12(5 − n) = 12(3 − p), respectively.
In the above background let us now follow the motion of another single parallel p-brane,
displaced from the original N by the radial coordinate-distance r. This motion is described by
the brane action, which can be decomposed as the sum of two pieces: the Born-Infeld (BI) and
the Wess-Zumino (WZ) parts. The dilaton and graviton couplings are given by the Born-Infeld
contribution, which is (in the String Frame):
Z q
SBI = −Tp dd ξ e−φ −det(gbM N ∂µ xM ∂ν xN + 2 `2s Fµν ) , (2.3)

where xM are the coordinates of the embedding, ξ µ are the world-volume coordinates and Tp
denotes the brane tension. Fµν denotes the field strength for any open-string gauge modes
confined to the brane, although for the moment we put these fields to zero.
The coupling to the bulk gauge field, A[p+1] , is given by the Wess-Zumino part of the brane
action: Z
SW Z = −q Tp A[p+1] , (2.4)
where q represents the brane charge, with q = 1 representing a probe brane and q = −1 repre-
senting a probe antibrane.
Since our branes are straight and parallel it is convenient to choose the following coordinate
gauge:
xµ = ξ µ , (2.5)
where as before xµ are the coordinates parallel to the world volume. To the extent that we follow
only the brane’s overall centre-of-mass motion – we return to the validity of this assumption in
the next consideration. We have seen that probe brane orbits decay due to radiation into bulk
modes, and but this radiation could also be accompanied by radiation onto the brane. We next

4
perform an estimate of this rate, along the lines of the one done earlier for bulk radiation, with
the conclusion that radiation into brane modes is subdominant to bulk radiation.
We use for these purposes the coupling of the brane-bound gauge field, given by the (String-
Frame) Born-Infeld action:
Z q
SBI = −Tp dd ξ e−φ −det(M N ) [1 + `4s µλνρ Fµν Fλρ + · · ·] , (2.6)

where µν = gbM N ∂µ xM ∂ν xN is the brane’s induced metric and Fµν is the gauge field strength. We
see that the relevant coupling is in this case κB = Tp `4s e−φb , where φb = φ(yb ) is the dilaton field
evaluated at the position of the brane.
The effective action of a (Dp, Dp) brane system in Type IIA(B) theory should be given by
some extension of the DBI action and the WZ terms which include the tachyon fields. The DBI
part may be given by the projection of the effective action of two non-BPS Dp -branes in Type
IIB(A) theory with (−1)FL projection. We are interested in this paper in the appearance of
tachyon, gauge field and the RR field in these actions. These fields appear in the DBI part as
the following:
Z  q 
SDBI = −Tp dp+1 σT r V (T ) −det(ηab + 2πα0 Fab + 2πα0 Da T Db T ) , (2.7)

where Tp is the p-brane tension. The trace in the above action should be completely symmetric
between all matrices of the form Fab , Da T , and individual T of the tachyon potential. These
matrices are
(1)
!
Fab 0 0 Da T 0 T
   
Fab = , Da T = , T = (2.8)
0
(2)
Fab (Da T )∗ 0 T∗ 0
(i) (i)
where Fab = ∂a Ab − ∂b A(i) (1) (2)
a and Da T = ∂a T − i(Aa − Aa )T . The tachyon potential which is
consistent with S-matrix element calculations has the following expansion:
1
V (T ) = 1 + πα0 m2 T 2 + (πα0 m2 T 2 )2 + · · ·
2
where m2 is the mass squared of tachyon, i.e.m2 = −1/(2α0 ). The above expansion is consistent
0 2 2
with the potential V (T ) = eπα m T which is the tachyon potential of BSFT. This action has the
following expansion:
πα0  (1)
!

0 2 2 ∗
LDBI = −2Tp − Tp (2πα ) m |T | + DT · (DT ) − F · F (1) + F (2) · F (2) + · · ·(2.9)
2
where dots refers to the terms which have more than two fields.
Using the expansion for the exponential term in the WZ action (6.123), one finds many
different terms. The terms which involve at most three open string fields are the following:
µp (2πα0 )C ∧ STr iF = µp (2πα0 )Cp−1 ∧ (F (1) − F (2) ) (2.10)
µp µ p
n o
(2πα0 )2 C ∧ STr iF ∧ iF = (2πα0 )2 Cp−3 ∧ F (1) ∧ F (1) − F (2) ∧ F (2)
2! 2! n o
+Cp−1 ∧ −2β 2 |T |2 (F (1) − F (2) ) + 2iβ 2 DT ∧ (DT )∗
µp µp n o
(2πα0 )3 C ∧ STr iF ∧ iF ∧ iF = (2πα0 )3 Cp−3 3iβ 2 (F (1) + F (2) ) ∧ DT ∧ (DT )∗
3! 3!
The coupling of one RR field Cp−1 , two tachyons and one gauge field in the above terms can be
combined into the following form:
Z n o
0 2
2
−β µp (2πα ) C(p−1) ∧ d(A(1) − A(2) )T T ∗ − (A(1) − A(2) )d(T T ∗ )
Σ(p+1)
Z
= −β 2 µp (2πα0 )2 H(p) ∧ (A(1) − A(2) )T T ∗ (2.11)
Σp+1

5
 We want to create the higher dimensional model which captures the important features of
the formation of defects with bulk fields included, and at the same time is amenable to a lattice
regularization. This will allow us to follow the evolution to see whether defects form, and how
the interactions with the bulk fields affect that formation. We want to include the minimal field
content that will allow us to study the formation of the defects, so we will not include the metric
and dilaton fields present in the full 10-dimensional model. We can extend the extremal model
we used to a higher dimensional one which will have fewer differences with respect to the full
superstring theory model. The most straightforward generalization would be to consider a 3 + 1-
dimensional brane and a 9 + 1 dimensional bulk. The field content of the worldvolume theory
would still be the Abelian Higgs model, but the bulk field will now be a rank-2 antisymmetric
tensor field, which has the appropriate rank to couple to the 2-dimensional world-volume of a
string. The Abelian Higgs model in 3 + 1 dimensions admits stable string-like defects which in
our model will be charged under the bulk field. The Lagrangian of the model is:
" # " #
Z
3 1 2 µ ∗ ccs Z Z
9 1 2
L= d x dt − 2 F − Dµ φD φ − V (φ) − F ∧C + d x dt − 2
H ,
M4 4gbrane 2 M4 M10 12gbulk
(2.12)
where as before we denote by F = dA the field strength of the field A and by H = dC the field
strength of the field C. The equations of motion now become:
ναβγ
∂µ F µν + igbrane
2
(φDν φ∗ − φ∗ Dν φ) + ccs gbrane
2
Hαβγ = 0 , (2.13)
2
αβνλ
∂µ H µνλ − ccs gbulk
2
Fαβ δ (z) = 0 . (2.14)
2
If we now want to study how the solution for a Nielsen-Olesen string is modified by the
presence of the bulk field, for the brane fields we make the usual ansatz for a string placed
along the z-direction: far from the string the scalar and gauge fields. Since we want to study
the formation of codimension 2 defects the brane must have at least 2 spatial dimensions, so
we will choose a 2+1 dimensional brane. The defects that form are 0+1-dimensional and the
corresponding bulk field that couples to their worldvolume must carry a single index. Therefore
we will have a vector field living in the bulk. This field corresponds to the Cp−1 RR field in the
10-dimensional model. We choose the bulk to have the minimal space dimensionality, 1 space
dimension more than the brane. Inside the brane we put the same field content as in the full
+ −
10-dimensional model, an Abelian gauge field corresponding to the linear combination Fµν − Fµν
and a complex scalar field corresponding to the complex tachyon T .
In Type II superstring theories, there are two kinds of branes: stable or BPS branes, which are
supersymmetric and charged, and unstable or non-BPS branes, which are non-supersymmetric
and uncharged. The p-dimensional unstable D (Dp) branes eventually decay into stable D(p − 1)
ones. This decay process is described by the dynamics of a tachyon field T . The effective
action of this unstable brane system in low energy approximation is conjectured to be of the
Dirac-Born-Infeld (DBI) type form, as derived from superstring theory:
Z
√ q
S=− dp+1 x −gV (T ) 1 + ∂µ T ∂ µ T , (2.15)

where the effective potential V (T ) takes its maximal value at T = 0 and vanishes asymptotically
as T tends to infinity. The homogeneous decay of the tachyon field involves the field rolling to
its vacua at T = ±∞, towards a state which can be characterised as a pressureless fluid without
propagating modes. This is consistent with the notion that the open string states disappear from
the spectrum as the brane decays.
The equation of motion derived from the DBI action is:
V0
2T + (g µν ∂T · ∂T − ∂ µ T ∂ ν T )∂µ ∂ν T − (1 + ∂T · ∂T ) = 0. (2.16)
V

6
 We are interested in time-independent solutions, neglecting the gravitational and Ramond-
Ramond fields of the brane. It is known that there are solutions depending on only one space
coordinate, which we denote x.
The energy of this system is
Z q 1 Z p−1 Z
E= dp xV (T ) 1 + (∂x T )2 = d x dxV 2 . (2.17)
V0
The energy per unit (p − 1)-dimensional volume in one period of the tachyon field is σ1 = πVm /β
with the specific potential adopted in this paper. The tension of the parent Dp-brane is Tp = Vm ,
so this is the correct tension for a Dp − 1-brane, Tp−1 .
We will briefly recall the form of the effective TDBI action obtained in superstrings that we
introduced earlier. We will first see the case of the non-BPS brane then that of the (Dp, Dp)
brane systems. We will add the contribution of Ramond-Ramond fields in the Wess-Zumino term.
We will start with branes of maximum dimension, of dimension 9 + 1 because the expressions are
simpler and capture the physics of any other brane. Indeed, the actions on the lower dimensional
branes are obtained by T-duality along the directions made transverse. For the non-BPS brane,
the effective action of a D9-brane in type IIA superstring theory is

√ Z q
SnonBP S = 2T9 d10 σ eΦ V (T ) detGab + Bab + 2πα0 Fab + ∂a T ∂b T + SW Z (2.18)

The expression is abelian, since there is only one brane. We discuss its scope in chapter
refcc chap: mot. It extends apriori only along spatial condensations. To discuss the temporal
condensations the basis of study is the tachyon action in the static gauge :
√ Z
p+1
q
ST = 2Tp d σ V (T ) ηab + ∂a T ∂b T (2.19)
the Wess-Zumino term is known to be of the form
Z
0
C(m) ∧ eB+2πα F
X
S W Z = µp W (T )dT ∧ (2.20)
p+1 m∈ IIA

where W (T ) ∝ V (T ) and B and F respectively the pull-backs on the volume of the brane of the
Kalb-Ramond field and the Maxwell tensor of the open superstring gauge field. The charge µp is
proportional to the voltage of the brane. The R-R gauge fields are type IIA differential forms of
even index. In type IIB, they have an odd index. Note that the tachyon is not minimally coupled
to the gauge fields of the open strings, which requires their confinement during condensation.

3 The Dynamics in BSFT of the (Dp, Dp) Brane System

In superstring theory a pair of parallel brane-antibrane system constitutes an unstable object in
the theoretical construction. To study the dynamics of unstable D-branes, the BSFT is a useful
tool and it has provided a good understanding of tachyon condensation at the classical level. It
describes the off-shell dynamics of open strings in a fixed on-shell background of closed strings
in which an open string field configuration corresponds to a boundary term in the world-sheet
action of the string. Therefore, specifying a boundary term means giving the background values
of the various modes of the open string. It is based on the Batalin-Vilkovisky formalism whose
master equation provides the effective action of the theory. In the bosonic string theory, the disk
partition function of the open string theory Z and the BSFT action are related by the master
equation !
i ∂
S = 1+β Z (3.21)
∂g i

7
where g i are the couplings of the boundary interactions and β i are the corresponding world-
sheet β-functions. Given a specific form of the tachyon profile, the BSFT action reduces to the
effective action for the tachyon field allowing us to compute the tree level tachyon potential. For
superstrings, the tachyon β-function is zero and (3.21) reduces to

S=Z (3.22)

The partition function Z was computed in current research article. The disc partition function
is formally defined as Z
Z = DXDψ e−(Sbulk +Sbndy ) (3.23)
where
1 Z 2 2
 
µ
Sbulk = dz 0
∂X µ ∂Xµ + ψ µ ∂ψµ + ψ ∂ψ µ (3.24)
4π α
is the bulk action for the NSR superstring.
The boundary term of the system is computed introducing auxiliary boundary fermion su-
perfields ΓI = η I + θF I where I = 1, 2m , and, N = 2m−1 is the number of pairs. Consider, for
example, the case where we have 2m branes. The 2m × 2m matrices of the gauge group U (2m ),
generated by the branes, can be expanded in terms of SO(2m) gamma matrices. Now, instead
of gamma matrices, one can introduce 2m boundary fermion superfields ΓI with action
Z
1
S=− dτ dθ ΓI DΓI (3.25)
4
and, after canonically quantizing, one arrives at the anti-commutation relations {η I , η J } = 2δ IJ .
Thus, η I can represent the Clifford algebra needed for the expansion of the 2m × 2m matrices.
In the case of a single brane-antibrane system, expanding the resulting action in terms of the
component fields one has
Z "
α0 I I 1 I I α0 i µ 
Sbnry = − − T T + η̇ η + Dµ T I ψ µ η I + Ẋ Aµ + 12 Fµν ψ µ ψ ν
4 4 2 2
i   
1 0 IJ µ ν
+ Ẋ µ AIJ
µ + 2 α Fµν ψ ψ η I η J dτ (3.26)
4
Here I, J = 1, 2,
 
A±
µ =
1
2
Aµ ± iA12
µ

(3.27)
Dµ T I = ∂µ T I − iAIJ
µ T
J
(3.28)

and the gauge fields A± µ on the brane and anti-brane, respectively, have been expressed in terms
of the abelian gauge fields AIJ µ , (anti-symmetrized in I, J) and Aµ . Moreover, Fµν = ∂[µ Aν] and
F IJ µν = ∂[µ AIJ ν] . In the case of a constant tachyon field and zero gauge fields, the boundary
action reduces to
α0 Z
Sbdry = dτ T I T I (3.29)
4
Since there is no other dependence on the tachyon field in the bulk action, we learn that in this
case the tachyon potential for the system is
0 2
V0 (T ) = 2T9 e−2πα |T | (3.30)

where we defined T = 12 (T 1 + iT 2 ), whereas T9 denotes the tension of a D9 -brane which is

defined, for general p, by
1
Tp = p+1 (3.31)
p
(2π) 2 gs

8
where gs is the string coupling constant. The stable vacua is at T = ∞, where the vacua
energy vanishes. Since the potential (3.30) is exact, it gives a proof of Sen’s conjecture that
the negative energy contribution from the tachyon precisely cancels the D-brane tension: under
tachyon condensation, the D-brane will decay into the closed string vacuum without any D-
branes, therefore, excitations are described by closed strings alone.
Let us turn now to the case of a spatially dependent tachyon. In this case, by a combination
of spacetime and gauge rotations one can bring T I to the following form:
√
α 0 T I = uI X I (3.32)

where uI are constants. When the gauge fields are zero, one can compute the partition function
(3.23) using eqs. (3.24), (3.26) and (3.32). The result is [?]
Z 2
0
d10 X0 e−2πα T T F (πα02 (∂I T I )2 )
Y
Z = 2T9 (3.33)
I=1

where
4x xΓ(x)2
F (x) = (3.34)
2Γ(2x)
The partition function allows us to have an expression for the action of the tachyon at all orders
in derivatives. However, there is an ambiguity in the expansion, because any term with at least
two derivatives acting on T can be added. At quadratic order the result is unambiguous
Z
0
h i
S ≈ 2T9 d10 x e−2πα T T 1 + 8πα02 ln(2) ∂ µ T ∂µ T + . . . (3.35)

where has been used the expansion

F (x) = 1 + 2ln(2) x + O(x2 ), x→0 (3.36)

Now, consider the case where one of the spatial directions, y, is wrapped on a circle of radius
Re ≤ √ and that we have a constant Wilson line A wrapping the compact direction on say the D
9
brane. The gauge field strength in (3.26) vanishes and the only dependence on the gauge field
comes from the covariant derivative. We can lift the above expression to include the covariant
derivative by simply changing the argument of the function F .
Applying a T-duality transformation along y, the gauge field is mapped to the Higgs field which
measures the distance d between a pair, separated along the dual coordinate ye with d ∼ |A|.
Adopting the normalization of the tachyon field the action (3.35) becomes
Z h i
2
S = 2T9 d9 x dy e−|T | 1 + 2α0 |∂µ T |2 + 2α0 A2 |T |2 (3.37)

The potential term is h i

2
V0 (T ) = 2T9 e−|T | 1 + 2α0 A2 |T |2 (3.38)
The extrema are given by
∂V0 (T ) 2
  
= 2T9 |T |e−|T | 4A2 α0 − 2 2A2 α0 |T |2 + 1 = 0 (3.39)
∂|T |
i.e., √
2A2 α0 − 1
|T | = 0, |T | = +∞, and |T | = √ 0 (3.40)
2α A
To study the nature of these extrema, we need to compute the second derivative: around |T | = 0
we have
∂ 2 V0 (T ) 2

0 2

||T |=0 = m = 4T9 2α A − 1 (3.41)
∂ 2 |T |

9
 q
1
Therefore, we see that this potential has a minimum at |T | = 0 if A > 2α0
or it has a true
q
1
tachyonic instability if A < .
This behavior has a clear physical interpretation: recall that
2α0
our model is equivalent to the case of a parallel brane-antibrane system separated by a distance
d. If the distance d is large enough, then the tachyon mode between the two should go away,
since the tachyon field comes from the open string suspended between the two branes and thus
that string acquires a mass lift when two branes are distant. Notice also that in order to get
a canonical kinetic term in the BSFT action we must perform the following redefinition of the
tachyon field: T = T (φ) with
q Z |T |
2 /2
φ= 8α0 T9 ds e−s (3.42)
0
With this redefinition, the action (3.37) becomes
Z  
S= d9 x dy 1
2
(∂φ)2 + V0 (T (φ)) (3.43)

and the tachyon vacuum at infinity is placed at a finite value of the new field φ. Indeed, the two
local minima are q
φ0 = 0 , φ1 = 4πα0 T9 . (3.44)
This redefinition allows us to compute the mass of the tachyon: in the presence of a Wilson line
A it is given by
∂ 2 V (φ) 1 |T |2 − 1
" #
 
2 0 2 4 2
M = = + α A |T | − 4|T | + 1 (3.45)
∂φ2 α0 2
whereas if A = 0 we have
2 1 |T |2 − 1
MA=0 = (3.46)
α0 2
Notice that the same results were found in the current literature but with different methods.
Henceforth, we will consider only the real part of the tachyon field: this is consistent with
the tachyon equations of motion and it is also a natural setup since we are not interested in
lower dimensional D-brane left after the tachyon condensation which needs complex tachyon
configurations.
The N = (1, 1) superspace action on the disk was written in research article, including the
coupling to background gauge and tachyon fields. In the present context one considers non-trivial
Wilson Lines along the circle, T-dual to the brane positions x1 and x2 along X, the T-dual of
Y . They naturally appear in the form x(±) = x1 ± x2 . Setting aside the ’spectator’ dimensions,
one considers a pair of N = (1, 1) superfields on the disk, one time-like (X0 ) and the other com-
pactified on a circle (Y ), with e.g. X0 = X0 + √i2 (θψ0 + θψ 0 ) + θθF0 . The superspace coordinates
are denoted as zb = (z, θ, θ). At the boundary of the disk, the Grassmann coordinates satisfy the
boundary condition θ = ±θ. The algebra of the Chan-Patton factors for the brane-antibrane
system is conveniently implemented by the canonical quantization of boundary fermions [?], see
below. These boundary fermions are the bottom components of fermionic superfields of the
boundary N = 1 superspace. For the brane-antibrane system one needs a complex superfield
Γ± = η ± + θF ± . (3.47)
with Γ− = (Γ+ )∗ .
Then the worldsheet action on the disk, including the tachyon background as well as Wilson
lines around the circle, reads:
1 Z I
x(+)
SBCF T (λ+ , λ− ) = d2 z d2 θ −DX 0 DX 0 + DY DY + i du dθ Du Y
2π D2 S1 4π
I
x(−)
− du dθΓ+ Du + i Du Y Γ− − Γ+ T + − Γ− T − , (3.48)
S1 2π

10
with the measure 2 θ = θ θ, the superspace holomorphic derivative D = ∂θ +θ∂ and the superspace
boundary derivative Du = ∂θ + θ∂u , with the boundary coordinate u on S 1 .
We consider simple rolling tachyon profiles of the form:

λ± ωX 0
T± = e , (3.49)
2π
√
with 0 < ω1 2. In order to get a real action, one chooses (λ+ )∗ = λ− . These are actually the
tachyons that we are expecting to be solutions of the spacetime effective action. It is understood
in this expression that the superfield X is taken on the superboundary of the disk.
(−)
The space-time gauge field A(−) = − x4π dy being locally pure gauge, its minimal coupling to
the fermionic superfields can be absorbed by the gauge transformation. One has to be careful
with this transformation if Y -dependent insertions appear in the path-integral; a prescription
must be chosen.
x(−)
Γ± → Γ± e±i 2π Y . (3.50)
After this field redefinition, the boundary fermionic superfields are free, with the propagator on
the real axis:
Γ+ (zb)Γ− (w)
b = b(zb − w)
b = (z − w) − 2 θz θw δ(z − w) , (3.51)
with the sign function (z) = Θ(z) − Θ(−z). This implies that ∆(Γ± ) = 0, i.e. vanishing
conformal dimension.
In terms of these new variables the worldsheet action reads:
1 Z I
x(+)
SBCF T (λ+ , λ− ) = d2 z 2 θ −DX 0 DX 0 + DY DY + i du θ Du Y
2πI D2 S1 4π
− du θ Γ+ DΓ− − Γ+ T + − Γ− T − (3.52)
S1

where the tachyon fields have now the expression:

λ± ±i x(−) Y +ωX 0
T± = e 2π . (3.53)
2π

Starting from the action eq:action, renaming Y as X,

f and integrating over the fermionic
coordinates one gets the action:

+ − 1 Z 2 0 0
I
x(+) f
SBCF T (λ , λ ) = d z −∂X ∂X + ∂X∂X + i du ∂u X
2π D2 S1 4π
I
λ+ + + + λ− − − −
+ duη + ∂u η − − η ψ T − η ψ T
IS
1 2π 2π
− duF + F − − F + T + − F − T − , (3.54)
S1

with:
√ √
ψ ± = ±ir 2ψex + ω 2ψ 0
0
T± = e±irX+ωX
e
. (3.55)

Auxiliary fields F ± are then integrated to give:

+ − 1 Z 2 0 0
I
x(+) f
SBCF T (λ , λ ) = d z −∂X ∂X + ∂X∂X + i du ∂u X
2π D2 S1 4π
I
λ+ + + + λ− − − − + −
2λ λ
+ duη + ∂u η − − η ψ T − η ψ T + ε1−4r T + T −(3.56)
.
S1 2π 2π 4π 2

11
The contact term at the end of second line shows up, with a UV cutoff ε. This term, that
does not follow from √the equations of motion contributes nevertheless to correlation functions
when 1/2 < |r| < 1/ 2. Finally, as the center-of-mass perturbation completely factorizes and
commutes with any operators, one can set x(+) = 0 without loss of generality.
Upon quantizing canonically the boundary fermions η ± , one recovers the Chan-Patton algebra
corresponding to the brane-antibrane system, where now the prescription for the path integral is
Z
i ,ψ i ]
Z = Tr DX i Dψ i Pe−S[X (3.57)

which includes a path ordering for the operator insertions and a trace over the CP factors. In this
context the tachyon becomes a boundary changing operator and when inserted on the boundary
of the disk, it interpolates between the two distinct boundary conditions corresponding to the
brane and to the antibrane.
The worldsheet action on the disk takes finally the form
I
λ+ + 0 λ− − 0 λ+ λ− 1−4r2 2ωX 0
S = Sbulk − du σ ⊗ ψ + eirX+ωX + σ ⊗ ψ − e−irX+ωX − ε e . (3.58)
e e
S1 2π 2π 4π 2
Coming back to the brane-antibrane system, we consider the following worldsheet action on
the upper half-plane, as a function of the boundary couplings. So now we take the boundary
variable to be u ∈ R. For convenience, we rescale the coupling according to λ± → 2πλ± .
Z  0 0 δr 3 
S = Sbulk − dx λ+ σ + ⊗ ψ + eirX+ωX + λ− σ − ⊗ ψ − e−irX+ωX − i σ ⊗ ∂u X (3.59)
e e f
2
We omitted for the moment the contact term, which will enter later on in the discussion.
In BSFT of open strings we have a relation between the action of BSFT - on the space of
field theories - on-shell and the partition function calculated on the disk D2 which in supersym-
metric theory appears to be particularly simple apriori if we assume that matter and ghosts are
decoupled:

S[φion ] = ZD2 [φion ] (3.60)

where the expression of the partition function is explicitly
Z P H
−Sbulk [Gab ,Bab ,Φ]− φion Vi
ZD2 [φion ] = tr P [dX][dψ]e i S1 (3.61)

In this context, φion is a constant value corresponding to a fixed point of the renormalization
group for the coupling associated with the vertex operator Vi . The equality suggests by extraction
of the zero mode of the bosonic fields X a that we can express an action density, the lagrangian
as a function of the partition function density noted Z 0 according to:
Z Z
p+1 0
d x L[ϕion (x)] = dp+1 x ZD i
2 [ϕon (x)] (3.62)

The expression of ϕion (x) is simply given by the identity ϕion (xa ) = φion Vi (X a )0 with . . .0 the
correlator calculated in free theory. Therefore, if we want to obtain the on-shell action along a
rolling tachyon which we know to be a solution of the equations of motion in the case of a only
non-BPS brane or a pair of separate brane-antibrane system, we must calculate the partition
function on the disk for which we add a supersymmetric deformation:
I
δS = T (X 0 ) (3.63)
S1
We are therefore looking at a fixed distance brane-antibrane system which we have shown to
be for all r a BCFT. We need to calculate the partition function for this system and we will start
from the action defined on the following disk superspace:

12
 I −
I
λ+ + irX+ωX
I
0 λ − −irX+ωX 0
S = Sbulk − Γ+ DΓ− − i
Γ e −i Γ e (3.64)
e e
2π 2π
Here the tachyon is therefore chosen on-shell at a constant distance. In the bulk the back-
ground is trivial, that the space is flat and that Bab = 0 and Φ = φ constant. On the unit disk
we will use that the variables are z = ρeit with ρ < 1 and in particular on the edge, so along the
unit circle we will have z = eit . So the formula we use to calculate the partition function is:

Z H
Γ+ DΓ− −i
H dzdθ λ+ Γ+ eirX+ωX 0 H dzdθ λ− Γ− e−irX+ωX 0
ZD2 [λ± , r] = [dΓ+ dΓ− ][dX][dX 0 ]e−Sbulk − −i
e e
S1 2π S1 2π

(3.65)
Nevertheless, there have been several attempts to generalize the BSFT to the one-loop am-
plitude in the (Dp, Dp) brane systems. All of them assume that the relation (3.22) is still true at
one loop. Then, they construct the partition function at one loop by keeping fixed the boundary
of the disk and the tachyon profile on it and adding more boundaries and handles to the string
world-sheet diagram. In particular, on the annulus one has
Z
S[U] = Z[Af ixed , B, U] (3.66)
annulus

where U is the coefficient of the linear tachyon profile (3.32), Af ixed is the boundary of the
disk and B is the inner boundary of the annulus. Similarly, the cylinder amplitude can be
computed in the closed string channel using the boundary state formalism. It is well known
that the partition functions obtained in the two different schemes agree on-shell thanks to the
open-closed string duality. However, the presence of the tachyon takes the theory off-shell and
it is not clear, a priori, that the two different schemes yield the same result. In particular, since
the boundary interactions are due to non-primary fields, the use of conformal maps to transform
one worldsheet into another one is not helpful because the transformation laws of the fields are
unknown. However, it seems that at one-loop at least, the two results are equivalent.

4 World-volume Fermions and κ-symmetry

For definiteness we shall restrict our analysis to D-branes in type IIA superstring theory, but
generalization to type IIB theory is straightforward following the analysis of in the current lit-
erature. On a non-BPS Dp-brane world-volume in type IIA superstring theory, we have a 32
component anti-commuting field Θ which transforms as a Majorana spinor of the 10 dimensional
Lorentz group. We shall denote by ΓM the ten dimensional γ-matrices, and take the indices
M, N to run from 0 to 9. In order to construct the world-volume action involving the fields Aµ ,
Y I , Θ and T (0 ≤ µ ≤ p, (p + 1) ≤ I ≤ 9) in static gauge, we first define:

Πνµ = δµν − ΘΓν φµ Θ, ΠIµ = φµ Y I − ΘΓI φµ Θ, (4.67)

Gµν = ηM N ΠM N
µ Πν + φµ T φν T , (4.68)
and
h 1 i
Fµν = Fµν − {ΘΓ11 Γν φµ Θ + ΘΓ11 ΓI φµ Θφν Y I − ΘΓ11 ΓM φµ ΘΘΓM φν Θ} − {µ ↔ ν} , (4.69)
2
where
Fµν = φµ Aν − φν Aµ . (4.70)
In terms of these variables, the DBI part of the world-volume action is given by
Z q
SDBI = − dp+1 x V (T ) −det(G + F) . (4.71)

13
The action is invariant under the supersymmetry transformation parametrized by a ten dimen-
sional Majorana spinor . In the static gauge in which we are working, the infinitesimal super-
symmetry transformation laws are given by

δp Θ =  − (Γµ Θ)φµ Θ, δp Y I = ΓI Θ − (Γµ Θ)φµ Y I , δp T = −(Γµ Θ)φµ T ,

1
δp Aν = Γ11 Γν Θ + Γ11 ΓI Θφν Y I − (Γ11 ΓM Θ ΘΓM φν Θ + ΓM Θ ΘΓ11 ΓM φν Θ)
6
− (Γµ Θ)φµ Aν − φν (Γµ Θ)Aµ . (4.72)

The subscript p in δp denotes that these are the supersymmetry transformation laws on the D-
p-brane world-volume. The supersymmetry transformation parameter  is a Majorana spinor of
the ten dimensional Lorentz group.
Besides the DBI term, the world-volume action also contains a Wess-Zumino term. In the
bosonic sector this term is important only for non-vanishing RR background field, but once we
take into account the world-volume fermions, this term survives even for zero RR background.
The structure of this term is
Z
SW Z = W (T ) dT ∧ C ∧ eF , (4.73)

where F = Fµν dxµ ∧ dxν , W (T ) is an even function of T which vanishes as T → ±∞, and
C is a specific combination of background RR fields and the world-volume fields Y I , Θ on the
D-brane[2]. In particular, the bosonic part of C is given by q≥0 C (p−2q) where C (p−2q) denotes
P

the pull-back of the RR (p − 2q)-form field on the D-p-brane world-volume. This vanishes for
vanishing RR background, but there is a part of C involving the world-volume fermion fields
that survives even in the absence of any RR background. Since we shall not need the explicit
form of C for our analysis, we shall not give it here. The Wess-Zumino term is also invariant
under the supersymmetry transformations (4.72). Later we shall see that consistency requires:
Z ∞ Z ∞
W (T )dT = V (T )dT = Tp−1 , (4.74)
−∞ −∞

Since we want to compare the world-volume action on a kink solution with that on the BPS
D-(p − 1)-brane, we need to first know the form of the world-volume action on a BPS D-(p − 1)-
brane. The world-volume fields in this case consist of a vector field aα (ξ) (0 ≤ α ≤ (p − 1)), a
set of (9 − p + 1) scalar fields which we shall denote by y I (ξ) ((p + 1) ≤ I ≤ 9) and y p (ξ) ≡ t(ξ)
respectively in the convention, and a Majorana spinor θ(ξ) of the ten dimensional Lorentz group.
Here {ξ α } denote the world-volume coordinate on the D-(p − 1)-brane. The DBI part of the
action is given by Z q
Sdbi = −Tp−1 dp ξ −det(g + f ) , (4.75)
where
gαβ = ηM N παM πβN , (4.76)
παβ = δαβ − θΓβ φα θ, παI = φα y I − θΓI φα θ, παp = φα t − θΓp φα θ, (4.77)
h 1 i
fαβ = − {θΓ11 Γβ φα θ + θΓ11 ΓI φα θφβ y I + θΓ11 Γp φα θφβ t − θΓ11 ΓM φα θθΓM φβ θ} , (4.78)
2
The Wess-Zumino term, on the other hand, has the form:
Z
Swz = Tp−1 c ∧ ef , (4.79)

where f = fαβ dξ α ∧ dξ β , and c is an expression containing the RR background and the world-
volume fields y I , t, θ. The bosonic part of c is given by q≥0 C (p−2q) where C (p−2q) denotes the
P

pull-back of the RR (p − 2q)-form field on the D-(p − 1)-brane world-volume. Like C, c also

14
contains a term involving y I and θ which survive even for trivial RR background. If we think of
the world-volume of the D-(p − 1)-brane as sitting inside that of a D-p-brane along the surface
xp = t(ξ), then c is in fact the pullback of C appearing in (4.73) provided we identify θ and y I
as the restriction of Θ and Y I along the surface xp = t(ξ).
Both Sdbi and Swz are separately invariant under the infinitesimal supersymmetry transfor-
mation:
δp−1 θ =  − (Γα θ)φα θ, δp−1 y I = ΓI θ − (Γα θ)φα y I , δp−1 t = Γp θ − (Γα θ)φα t ,
1
δp−1 aβ = Γ11 Γβ θ + Γ11 ΓI θφβ y I + Γ11 Γp θφβ t − (Γ11 ΓM θ θΓM φβ θ + ΓM θ θΓ11 ΓM φβ θ)
6
− (Γα θ)φα aβ − φβ (Γα θ)aα . (4.80)
The subscript (p − 1) on δp−1 indicates that these represent supersymmetry transformation laws
on the world-volume of a BPS D-(p − 1)-brane.
In order to show that the world-volume action Sdbi + Swz on the BPS D-(p − 1)-brane arises
from the world-volume action on the tachyon kink solution, we need to first propose an ansatz
relating the fields T (x, ξ), Aµ (x, ξ), Y I (x, ξ) and Θ(x, ξ) to the fields aα (ξ), y I (ξ), t(ξ) and θ(ξ)
on the BPS D-brane. For this we propose the following ansatz:
 
T (x, ξ) = f a(x − t(ξ)) , Y I (x, ξ) = y I (ξ), Θ(x, ξ) = θ(ξ),
Ax (x, ξ) = 0 Aα (x, ξ) = aα (ξ) .
(4.81)
We can now compute Gµν and Fµν in terms of the variables aα , y I , t and θ using (4.67)-(4.70)
and (4.81). The result is:
Gxx = 1 + a2 (f 0 )2 , Gαx = Gxα = −a2 (f 0 )2 φα t − θΓp φα θ ,
Gαβ = gαβ + φα t θΓp φβ θ + φβ t θΓp φα θ + (a2 (f 0 )2 − 1)φα tφβ t ,
Fαx = −Fxα = −θΓ11 Γp φα θ ,
Fαβ = fαβ − φα t θΓ11 Γp φβ θ + φβ t θΓ11 Γp φα θ , (4.82)
with gαβ and fαβ defined as in (4.76). Using manipulations we can now show that
det(G + F) = a2 (f 0 )2 {det(g + f ) + O(a−2 )} , (4.83)
and
Z q Z q
p+1
SDBI = − d x V (T ) −det(G + F) = −Tp−1 dp ξ −det(g + f ) = Sdbi . (4.84)

The analysis for SW Z is even simpler; – indeed this term was designed to reproduce the Wess-
Zumino term on the world-volume of a kink solution. For this let us define u = x − t(ξ) Then
from (4.82) we get
F ≡ Fµν dxµ ∧ dxν = 2Fxβ dx ∧ dξ β + Fαβ dξ α ∧ dξ β
= 2θΓ11 Γp φα θdu ∧ dξ α + fαβ dξ α ∧ dξ β . (4.85)
Since we have dT = af 0 (au) du only the second term on the right hand side of (4.85) will
contribute to SW Z given in (4.73). Thus we can replace F by f in (4.73). On the other hand, we
can analyze C by writing it as
C(q) µ1
∧ · · · dxµq
X
C = µ1 ···µq dx
q

(q C(q) α2
∧ · · · dξ αq + C(q) α1
∧ · · · dξ αq )
X
= xα2 ···αq dx ∧ dξ α1 ···αq dξ
q
Xh i
= q C(q)
xα2 ···αq du ∧ dξ
α2
∧ · · · dξ αq + (q C(q) (q)
xα2 ···αq φα1 t + Cα1 ···αq ) dξ
α1
∧ · · · dξ αq ,
q
(4.86)

15
where in the last step we have used dx = du + φα tdξ α . The term proportional to du does not
contribute to (4.73), whereas the term proportional to dξ α1 ∧ · · · dξ αq , after being summer over
q, is precisely the pull-back of C on the kink world-volume along x = t(ξ) and hence can be
identified with c. Thus using (4.74) we get
Z Z
SW Z = W (f (au)) a f 0 (au) du ∧ c ∧ ef = Tp−1 c ∧ ef = Swz , (4.87)

This shows that SDBI + SW Z reduces to Sdbi + Swz under the identification (4.81). In principle
we also need to check that any solution of the equations of motion derived from Sdbi + Swz is
automatically a solution of the equations of motion derived from SDBI + SW Z . Presumably this
can be done following the structural analysis, but we have not worked out all the details.
Finally, we need to check that the supersymmetry transformations (4.80) are compatible with
the supersymmetry transformations (4.72). For this we need to calculate δp−1 Aµ , δp−1 Y I and
δp−1 T using (4.80), (4.81) and compare them with (4.72). The calculation is straightforward,
and we get:

δp Ax = δp−1 Ax + Γ11 Γp θ, δp Aα = δp−1 Aα − Γ11 Γp θφα t,

I I
δp Y = δp−1 Y , δp T = δp−1 T . (4.88)

Thus we see that δp and δp−1 differ for the transformation laws of Ax and Aα . This difference,
however, is precisely of the form induced by the function φ(x, ξ) with φ(x, ξ) = Γ11 Γp θ(ξ). As
was argued below, this is a gauge transformation. Thus we see that the action of δp and δp−1
differ by a gauge transformation in the world-volume theory on the D-p-brane.
This establishes that the world volume action on the kink reduces to that on a D-(p−1)-brane.
The latter has a local κ-symmetry which can be used to gauge away half of the world-volume
fermion superfields. This leads to a puzzle. Whereas on a BPS D-brane the local κ-symmetry
is postulated to be a gauge symmetry, i.e. different configurations related by κ-transformation
are identified, on a kink solution the appearance of the κ-symmetry seems accidental and a
priori there is no reason to identify field configurations which are related by κ-symmetry. We
believe the resolution of this puzzle lies in the general principle advocated below that any local
transformation of the fields which does not change the action must be a gauge symmetry. This
will automatically imply that the κ-transformation is a gauge transformation and we should
identify the configurations related by κ-transformation. This κ-symmetry can now be used to
gauge away half of the fermion fields on the world-volume of the kink.

5 The Tachyon Condensation with Bulk Fields

Study of various aspects of tachyon dynamics on a non-BPS D-brane of type IIA or IIB superstring
field theory has led to some understanding of the tachyon dynamics near the tachyon vacua. The
proposed tachyon effective action, describing the dynamics of the tachyon field on a non-BPS
Dp-brane of type IIA or IIB superstring theory, is given by:

Z
S = dp+1 x L ,
√
L = −V (T ) −detA , (5.89)

where
Aµν = ηµν + φµ T φν T + φµ Y I φν Y I + Fµν , (5.90)
Fµν = φµ Aν − φν Aµ . (5.91)
Aµ and Y I for 0 ≤ µ, ν ≤ p, (p + 1) ≤ I ≤ 9 are the gauge and the transverse scalar fields on the
world-volume of the non-BPS brane, and T is the tachyon field. V (T ) is the tachyon potential

16
which is symmetric under T → −T , has a maximum at T = 0, and has its minimum at T = ±∞
where it vanishes. We are using the convention where η = diag(−1, 1, . . . 1) and the fundamental
string tension has been set equal to (2π)−1 (i.e. α0 = 1).
The usual approach to defect formation during tachyon condensation is to take into account
only the fields that live in the worldvolume of the decaying non-BPS brane or brane-anti-brane
pair. This approach seems to be motivated by the fact that the resulting lower-dimensional branes
formed during the decay are localized inside the worldvolume of the parent brane. However, the
final state defects are themselves D-branes, and therefore couple to bulk RR-fields. One should
include the effects of these fields in the defect formation process.
When studying the formation of topological defects by the brane worldvolume fields only, one
usually considers a single non-BPS brane with the action:
Z q
S = −Tp dp+1 x e−φ V (T ) det (P [Gab + Bab ] + 2πα0 [Fab + ∂a T ∂b T ]) , (5.92)

where P [Gab + Bab ] represents the pull-back of the bulk metric and NS-NS two-form field on the
brane and Fab is the field strength of the Abelian world-volume gauge superfield. This action was
used extensively to study the evolution of the tachyon field in various special settings. The most
common one is the spatially uniform field in a Friedmann-Robertson-Walker universe. In this case
one usually sets the NS-NS two-form field and the brane gauge field to vanish everywhere, and
study the time evolution of the tachyon field and scale factor of the universe. The uniform field
does not lead to the formation of defects, it behaves like a pressureless fluid known as “tachyon
matter”.
When studying the formation of topological defects during tachyon condensation, one usually
chooses a flat metric, sets the NS-NS field and the brane gauge field to vanish everywhere, but
chooses a tachyon field that is both time and space dependent. The equation of motion for the
tachyon field
√ 
√ √
∂α −gg αβ ∂β T −gg αβ ∂β T ∂α (∂µ T ∂ µ T ) V 0 (T ) −g
√ α
− 3/2
− √ = 0, (5.93)
1 + ∂α T ∂ T α
2 (1 + ∂α T ∂ T ) V (T ) 1 + ∂α T ∂ α T

is non-linear, and one usually solves the equation around a point where T = 0. There the space
and time dependence can be approximated by a linear space profile with a time-dependent slope,
T (t, x) ' u (t) x, and the resulting equation for u (t) can be solved. The solution becomes singular
in finite time, the occurrence of the singularity marking the formation of the topological defect.
This result confirms the String Theory calculation in which a linear tachyon profile T (x) = ux
reproduces the correct tension of a codimension one brane in the limit u → ∞.
The brane gauge superfield is usually included in the form of a uniform background electric
field or a constant gauge potential (Wilson line). The case where both the brane gauge superfield
and a worldvolume scalar superfield other than the tachyon are included. This action is highly
non-linear and in order to perform a lattice regularization we prefer an expression in which the
square-root is expanded to quadratic order in the field strengths. Also for a single non-BPS
brane the tachyon is a real scalar which cannot be minimally coupled to the world-volume gauge
superfield.
It is therefore more convenient to consider the action for a brane-anti-brane pair. In the case
of a (D9, D9) brane system the expanded action involving the complex tachyon coupled to the
gauge superfields living inside each brane is given
Z "
−φ −2πα0 T T
S = 2TD9 10
d xe e 1 + 8πα0 ln (2) Dµ T Dµ T +

(2πα0 )2  + 2 (2πα0 )2  − 2 βα0 2  +

#
− 2

Fµν + Fµν + Fµν − Fµν . (5.94)
8 8 8

17
The two gauge fields live in the worldvolume of each brane and the tachyon field couples only
with one linear combination:  
−
Dµ T = ∂µ T − A+µ − A µ T . (5.95)
The brane fields couple with the Ramond-Ramond (RR) bulk fields through the Chern-Simons
coupling also given in Ref. [?]:
Z
0
DD
SRR = TD9 C ∧ Str e2πiα F . (5.96)

One can expand the exponential above, and the leading order coupling between the brane fields
and the bulk RR field involves the same linear combination of brane gauge fields that couples to
the tachyon: Z  
SRR = 2πα TD9 Cp−1 ∧ F + − F −
DD 0
(5.97)
−
We see that the orthogonal linear combination, A+ µ + Aµ does not couple to any other fields, so
we will drop it from the action. Including the kinetic terms for the RR field, the dilaton and the
metric, the 10-dimensional action describing the decay of the brane-anti-brane pair is:
1 Z √
" #

2
 1
S= 2 −G d10 x e−2φ R + 2 (∇φ) − F2
2κ10 2 (p)! p
(2π 2 + β) α0 2  +
Z   
0 − 2
−2TD9 d10 x e−φ e−2πα T T 1 + 8πα0 ln (2) Dµ T Dµ T + Fµν − Fµν , (5.98)
8
where Fp is the corresponding field strength for the potential Cp−1 , Fp = dCp−1 . The Cp−1 will
be the only field we consider here, we will not include the dilaton and the metric in the extremal
model we consider. Here we expand the Chern-Simons coupling of the brane and bulk fields to
second order in α0 where
9−p
(−i) 2
Cµ0 ...µp dxµ0 ∧ . . . ∧ dxµp .
X
C= (5.99)
p=odd (p + 1)!

The supertrace of the matrix is defined as,

!
1 0
Str M = T r M, (5.100)
0 −1

and F is the curvature of the superconnection, given by

!
iF + − T T DT
iF = − . (5.101)
DT iF − T T

Here we are interested in the particular case of codimension defects, so the relevant coupling
will be with the Cp−1 RR-field that couples with the defects. We want to keep both the tachyon
and the gauge fields non-zero, so we will expand the exponential inside the supertrace in powers
of α0 :
! ! !
2πiα0 F 1 0 0 1 0 iF + −T T DT
Str e = Tr + (2πα ) T r
0 −1 0 −1 DT iF − −T T
(2πα0 )2
! ! !
1 0 iF + −T T DT iF + −T T DT
+ Tr − ∧ −
2 0 −1 DT iF −T T DT iF −T T
+... h i
= (2πiα0 ) F + − F − + (5.102)
1
   
2
(2πα0 ) DT ∧ DT − iT T F + − F − − (F + ∧ F + − F − ∧ F − ) + . . . .
2

18
The terms of the type F ∧ F couple with Cp−3 and count only in the formation of codimension
defects. Therefore, in the case of a (D9, D9) brane system, the important couplings between the
brane and the bulk fields are
Z n h i h  io
2
TD9 −iC8 ∧ (2πiα0 ) F + − F − + (2πα0 ) DT ∧ DT − iT T F + − F − . (5.103)

Since we are interested in the simplest model which involves such a coupling, when we construct
the toy model we keep only the interaction that corresponds to the leading term in α0 ,
Z h i
0
(2πα ) TD9 C8 ∧ F + − F − . (5.104)

The easiest setup to study the formation of defects in the original theory is to consider the
action for a brane-anti-brane pair for only the tachyon field in flat space-time, with all the other
fields turned off, since the lower-dimensional branes are in fact vortices of the complex tachyon
field. The action is simply:
Z
0
h i
S = 2TD9 d10 X e−2πα T T 1 + 8πα0 ln (2) Dµ T Dµ T . (5.105)

One can also consider the action for lower-dimensional brane-anti-brane pairs. The equation of
motion derived from the acion above is:
T
∂µ ∂ µ T − 2πα0 T ∂µ T ∂ µ T + = 0. (5.106)
4 ln 2
As in the case of the real tachyon field, one can approximate the profile of the field with a linear
one as the vortex will form at the place where T = 0, and solve the resulting equation for the
slope of the profile. The resulting defect formed in the decay of a (Dp, Dp) brane system is a
Dp − 2 brane.
In order to undestand this we have to go back to the calculation of the space-time action
calculated on linear tachyon profiles and estimate the action in the limit of infinite slope. The
calculation was done and we reproduce the important points here. In the case of a (D9, D9)
brane pair the space-time action for a linear tachyon profile is
Z 2
10 −2πα0 T T
   
I
F πα0 y I
Y
S y = 2TD9 dX e (5.107)
I=1

√  2
where T I = uI X I / α0 and y I = uI . The function F has the expression and in the large
argument limit it takes the form: √
F (x) ' πx . (5.108)
Calculating the action on the profile y 1 → ∞ and y 2 → ∞ the authors of Ref. [?] obtain:
h i
π 2 2
10 − 2 y (X ) +y (X )
  Z 1 1 2 2    
S y I = 2TD9 dX e F πα0 y 1 F πα0 y 2
s s
Z
8 2 2 q 2 0 1q 2 0 2 2 0
Z
= 2TD9 dX π α y π α y → 4π α TD9 dX 8 . (5.109)
y1 y 2

 √ 2
The result gives the correct tension for a D7 brane, TD7 = 2π α0 TD9 . Regarding the RR
charge of the vortex, we can estimate it by using the result (5.102) in the expression of the
coupling between brane fields and the bulk RR fields,
Z
0 2 (2πα0 )2 Z
SRR = TD9 C8 ∧ e−2πα T T (2πα0 ) dT ∧ dT = TD9 C8 , (5.110)
α0

19
which again reproduces the correct result for the D7 brane RR charge.
These results allow us to obtain an upper limit for the density of defects formed, based only
on energetic considerations. The brane tension is equal to the mass per unit volume for the brane
and the result  √ 2
TDp−2 = 2π α0 TDp (5.111)
 √ 2
tells us that we can have at most one defect on each patch of area 2π α0 = (2πls )2 where
ls is the string length. This is a very large density and it shows that constraints other than the
energetic ones are more important in determining the final density of defects. In a realistic model
we expect that a very important role will be that of the other fields that we have neglected so far,
namely the dilaton and the graviton. These two fields have universally attractive interactions
and their presence allows for the existence of BPS states in which there is no interaction between
identical, parallel, branes. We expect the presence of these fields to also change the evolution
of the resulting network of defects, since our toy model allows for repulsive interactions between
same-charge defects, while no repulsive interactions are possible in the full Superstring Theory
model. The reduction of the number of fermionic degrees of freedom due to κ-symmetry for a
brane stretched in the T direction occurs here as well. The D-branes used in the construction of
are BPS in 9 + 1 dimensions. Thus, a D-brane localized on the circle transverse to the fivebranes
has a sixteen component fermionic field living on its worldvolume, while for a D-brane wrapping
the circle the analog of the κ-symmetry discussed above eliminates half of the fermions, and leaves
a dynamical eight component spinor on the brane. This agrees with spacetime expectations. It
is well known from the Hanany-Witten construction that a D-brane ending on a stack of N S5-
branes has the property that the massless fields on it are the worldvolume gauge field and fields
related to it by supersymmetry. Since the D/N S system preserves eight supercharges, those
fields form a vector multiplet of N = 1 supersymmetry in 5 + 1 dimensions, reduced to the p-
dimensional worldvolume of the D-brane. This multiplet contains an eight-component fermion,
in agreement with our discussion above.

6 The Effective Actions of (Dp, Dp) Brane Systems

In the scheme of effective field theory, instability of (Dp, Dp) brane-antibrane system is repre-
sented by a complex tachyon field (T, T ), and this tachyon is indispensable for the generation of
topological defects. Since this system possesses U(1)×U(1) gauge symmetry, we need two gauge
fields, Aaµ , a = 1, 2. Separation of the brane and the antibrane is described by scalar fields XaI
corresponding to the transverse coordinates of individual branes. When the U(1)×U(1) gauge
fields in (6.114) are rewritten as Aµ = (A1µ + A2µ )/2 and Cµ = (A1µ − A2µ )/2, the former Aµ
remains to be massless in symmetric phase and the latter Cµ becomes massive in broken phase.
It is easily read by the form of the action
Z hq q i
S = −Tp dp+1 x V (T ) +)+
−det(Xµν −) ,
−det(Xµν (6.112)

where Cµν = ∂µ Cν − ∂ν Cµ , Dµ T = (∂µ − 2iCµ )T , and

±
Xµν = gµν + Fµν ± Cµν + (Dµ T Dν T + Dν T Dµ T )/2. (6.113)
Note that the charge of T is 2 in the unit system of consideration. In this coincidence limit, the
D-brane is distinguished from the D-brane by coupling to the gauge field Cµ .
For Fµν , we employ the same configuration, i.e., all the components vanish for singular local
D-vortex and F0r is turned on for regular local D-vortex as has been done for global D-vortices.
Among various proposed tachyon effective actions, we shall deal with Dirac-Born-Infeld (DBI)
type effective action of the (Dp, Dp) brane system
Z
S = −Tp dp+1 x V (T, X1I − X2I )

20
 (r
h i
× 1 + ∂ X I ∂ X I + (D T D T + D T D T )/2
−det gµν + Fµν µ 1 ν 1 µ ν ν µ

r )
h i
2 + ∂ X I ∂ X I + (D T D T + D T D T )/2
+ −det gµν + Fµν (6.114)
,
µ 2 ν 2 µ ν ν µ

a
where Tp is the tension of the Dp-brane, Fµν = ∂µ Aaν − ∂ν Aaµ , and Dµ T = (∂µ − iA1µ + iA2µ )T . For
iχ
small tachyon amplitude τ from T = τ e , behavior of the tachyon potential is
" #
1
(T, X1I X2I ) (X1I − X2I )2 τ 2 + O(τ 4 ),
X
V − =1− − (6.115)
R2 I
√
where R is 2 in superstring theory.
In the context of DBI-type effective action describing D-brane systems with instability,
there have been much study on codimension-one solitons (codimension-one branes), particularly
tachyon kinks. For vortices (codimension-two branes), only the singular local vortex solution
with finite energy was constructed from the DBI action, and regular tachyon vortex solutions
were obtained in local field theory action with quadratic derivative terms and polynomial tachyon
potential.
Let us consider (Dp, Dp) brane system in the coincidence limit of two branes, X1I = X2I , with
fundamental strings. Then the macroscopic fundamental strings in fluid state are represented by
electric fluxes along their directions. Vortex-like codimension-two objects of our interest could
be interpreted as D(p − 2)-branes. For description of the global vortex-like objects, the gauge
fields and their field strengths should behave as A1µ = A2µ = Aµ and Fµν 1 2
= Fµν = Fµν in the
action (6.114). Then the action (6.114) in (1+p)-dimensions becomes
Z √
S = −2Tp dp+1 x V (τ ) −detX , (6.116)

where
Xµν = gµν + Fµν + (∂µ T ∂ν T + ∂ν T ∂µ T )/2. (6.117)
Additionally we assume that the produced D(p − 2)-branes are flat and all the transverse
degrees are frozen. Then we can neglect dependence of the transverse coordinates and it is
enough to find D0-branes from flat (Dp, Dp). In the context of solitons in the effective theory,
it is translated as point-like vortices on a plane. We will call this vortex as D-vortex in what
follows.
For unstable D-branes, the coupling to the bulk RR fields can be read off from the Wess-
Zumino term and, for (Dp, Dp), it is possibly be extended as
!
F 1 − T T i3/2 DT
Z
SWZ = µ Str CRR ∧ exp
Σ3 −i3/2 DT F 2 − T T
Z
= µ e−T T CRR ∧ (2C + iDT ∧ DT )
Σ3
d −τ 2 n
Z   
= 2µ CRR ∧ dr ∧ dθ e Cθ − , (6.118)
Σ3 dr 2
where µ is a real constant and the supertrace Str is defined to be a trace with insertion of
σ3 . Note that C ≡ Cµν dxµ ∧ dxν and Cµν = (Fµν 1 2
− Fµν )/2. For the (Dp, Dp) system, it is
obvious that all the singular and regular, global (Cθ = 0) and local D-vortices
 on the (Dp, Dp)
d −τ 2
Cθ − n2 ] = 2πµn) is
R
carry only D0 charge. In addition, the total D0 charge (= 4πµ dr dr [e
exactly proportional to the vorticity n (or the quantized magnetic flux (??) for the local vortex)
irrespective of the nature of the D-vortices, global or local, as expected.
The effective action of a (Dp, Dp) brane system in Type IIA(B) theory should be given by
some extension of the DBI action and the WZ terms which include the tachyon fields. The DBI

21
part may be given by the projection of the effective action of two non-BPS Dp -branes in Type
IIB(A) theory with (−1)FL . We are interested in this paper in the appearance of tachyon, gauge
field and the RR field in these actions. These fields appear in the DBI part as the following
Z  q 
SDBI = − dp+1 σT r V (T ) −det(ηab + 2πα0 Fab + 2πα0 Da T Db T ) , (6.119)

The trace in the above action should be completely symmetric between all matrices of the form
Fab , Da T , and individual T of the tachyon potential. These matrices are
(1)
!
Fab 0 0 Da T 0 T
   
Fab = , Da T = , T = (6.120)
0
(2)
Fab (Da T )∗ 0 T∗ 0
(i) (i)
where Fab = ∂a Ab − ∂b A(i) (1) (2)
a and Da T = ∂a T − i(Aa − Aa )T . If one uses ordinary trace, instead,
the above action reduces to the action was proposed after making the kinetic term symmetric and
performing the trace. This latter action is not consistent with S-matrix calculation. The tachyon
potential which is consistent with S-matrix element calculations has the following expansion:
1
V (|T |) = 1 + πα0 m2 |T |2 + (πα0 m2 |T |2 )2 + · · ·
2
where Tp is the p-brane tension , m2 is the mass squared of tachyon, i.e.m2 = −1/(2α0 ). The
0 2 2
above expansion is consistent with the potential V (|T |) = eπα m |T | which is the tachyon potential
of BSFT.
The terms of the above action which has contribution to the S-matrix element of one gauge
field and two tachyons in which we are interested in this paper are the following

πα0  (1)
!

0 ∗
LDBI 2 2
= −Tp (2πα ) m |T | + DT · (DT ) − (1)
F ·F +F ·F (2) (2)
+ Tp (πα0 )3
2
2
  
× DT · (DT )∗ F (1) · F (1) + F (1) · F (2) + F (2) · F (2) (6.121)
3
2m2 2  (1) 
+ |τ | F · F (1) + F (1) · F (2) + F (2) · F (2)
3
4 µ ∗ µ ∗

(1) µα (1) (1) µα (2) (2) µα (2)

− ((D T ) Dβ T + D T (Dβ T ) ) F Fαβ + F Fαβ + F Fαβ
3
Note that if one uses the on-shell value for the tachyon mass m2 = −1/(2α0 ), the above terms
would not be ordered in terms of power of α0 .
The WZ term describing the coupling of RR field to gauge superfield of brane-antibrane
system is given by
Z
0 0
 (1) (2)

S = µp C ∧ ei2πα F − ei2πα F , (6.122)
Σ(p+1)

where Σ(p+1) is the world volume and µp is the RR charge of the branes. In above equation, C
P p−m+1
is a formal sum of the RR potentials C = n (−i) 2 Cm . Note that the factors of i disappear
in each term of (6.122). The inclusion of the tachyon fields into this action has been proposed in
the literature using the superconnection of noncommutative geometry
Z
0
SW Z = µp C ∧ STr ei2πα F (6.123)
Σ(p+1)

where the curvature of the superconnection is defined as:

F = dA − iA ∧ A (6.124)

22
the superconnection is !
iA(1) βT ∗
iA = ,
βT iA(2)
√
where β is a normalization constant with dimension 1/ α0 which we shall find it later, and a
“supertrace” is defined by !
A B
STr = Tr A − Tr D .
C D
Using the multiplication rule of the supermatrices [?]
0 0
! ! !
A B A0 B 0 AA0 + (−)c BC 0 AB 0 + (−)d BD0
· = 0 0 (6.125)
C D C 0 D0 DC 0 + (−)a CA0 DD0 + (−)b CB 0

where x0 is 0 if X is an even form or 1 if X is an odd form, one finds that the curvature is
!
iF (1) − β 2 |T |2 β(DT )∗
iF = ,
βDT iF (2) − β 2 |T |2
(i)
where F (i) = 21 Fab dxa ∧ dxb and DT = [∂a T − i(A(1) (2) a
a − Aa )T ]dx . The WZ action (6.123) has
the following terms:

C ∧ STr iF = Cp−1 ∧ (F (1) − F (2) ) (6.126)

n o
C ∧ STr iF ∧ iF = Cp−3 ∧ F (1) ∧ F (1) − F (2) ∧ F (2)
n o
+Cp−1 ∧ −2β 2 |T |2 (F (1) − F (2) ) + 2iβ 2 DT ∧ (DT )∗
n o
C ∧ STr iF ∧ iF ∧ iF = Cp−5 ∧ F (1) ∧ F (1) ∧ F (1) − F (2) ∧ F (2) ∧ F (2)
n
+Cp−3 −3β 2 |T |2 (F (1) ∧ F (1) − F (2) ∧ F (2) )
o
+3iβ 2 (F (1) + F (2) ) ∧ DT ∧ (DT )∗
n o
+Cp−1 3β 4 |T |4 ∧ (F (1) − F (2) ) − 6iβ 4 |T |2 DT ∧ (DT )∗

The appearance of Cp−1 ∧ dT ∧ dT ∗ has been checked by studying the disk level S-matrix element
of one RR field and two tachyons. In the present paper we will check, among other things, the
appearance of Cp−1 ∧ DT ∧ (DT )∗ and Cp−1 ∧ |T |2 (F (1) − F (2) ) terms and fix their coefficients
using the S-matrix element of one RR field, two tachyons and one gauge superfield.
One can rewrite the supercurvature in complete form using the Clifford algebra. We replace
dx ...dxµn → n!1 γ µ1 ...γ µn , where γ µ satisfy the Clifford algebra {γ µ , γ ν } = 2g µν .
µ1

The supercurvature reads now

 1 µν 1
iγ µ Dµ T

γ Fµν − (T T − mm)
2
F= , (6.127)

 
1 µν 2
iγ µ Dµ T 2
γ Fµν − (T T − mm)

where γ µν = 21 [γ µ , γ ν ], namely dxµ ∧ dxν → γ µν . Note that in (6.127) we used the freedom
to add a constant part, represented by mm. In the non-commutative formalism with algebras
C ∞ (R4 ) ⊗ ( CI ⊕C I ), m, m correspond to the C I ⊕CI part.
There are two natural trace operations we can take over the Clifford algebra. We denote
by tr the one simply taken over the Clifford algebra elements, e.g. tr(γ µ γ ν ) = 2[(p+2)/2] g µν in a
(p + 1)-dimensional space. We denote by atr the antisymmetric trace over the Clifford algebra,
e.g. atr(γ µ γ ν ) = trγ µν , which leads naturally to the wedge-product structure. We denote by Tr
the one taken over the matrix structure of F.
Since the superconnection appears naturally in the description of the branes-antibranes system
it is natural to ask whether we can write the effective action in terms of the supercurvature. The

23
first hint is the (Dp, Dp) higher-dimensional effective action up to second order, as computed in
perturbative string theory
1 1
Z  
p+1
S2 = Tp d x F 1µν Fµν
1
+ F 2µν Fµν
2
− Dµ T Dµ T − (T T − mm)2 , (6.128)
4 4
where by Tp we denote the tension of a BPS Dp-brane. This action can be written as
Tp Z
S2 = − dp+1 xTr(trF 2 ) . (6.129)
2[(p+2)/2]
One may suspect then that the higher order terms in the effective action, in the slowly varying
fields approximation, where we neglect terms like ∂ k F and ∂ l T, l > 1, could be of the form
F n , n > 2. We will work in the slowly varying fields approximation in the following. We will see
in the next section that this approximation is sufficient for the analysis of some exact properties
of the tachyon condensation.
The second hint comes from the form of the Wess-Zumino (WZ) term of the branes-antibranes
system. It can be written as
Z
SW Z = τ dp+1 xTrs (atr(Γ C eF )) , (6.130)
e−mm µp
where τ is a normalisation constant, τ = 2[(p+2)/2]
and µp = gs Tp . Γ is given by

γe 0
 
p−1 p−1
Γ = i[ 2
]
, γe = i[ 2
]
γ 0 ...γ p , (6.131)
0 γe
and
1 µ1 ,···µn X
γ C= Cµ1 ,···µn (6.132)
n!
where Cµ1 ,···µn is an n-form corresponding to the RR n-form field.
In the language of differential forms (6.130) reads
Z
−mm
SW Z = µp e C ∧ Trs ( eF ) . (6.133)

This WZ action was proposed, and is expected in view the fact that D-branes charge is measured
by the K-theory class.
The supercurvature F can be decomposed as
 1 µν 1
iγ µ Dµ T
  
γ Fµν 1 0
2
F =  − (T T − mm) 
  
 
1 µν 2
iγ µ Dµ T 2
γ Fµν 0 1
= F − (T T − mm)11 . (6.134)

Using this form of the curvature the WZ action (6.133) can be written as
 
n
Z
F 
dp+1 xe−T T C ∧ Trs 
X
SW Z = µp . (6.135)
n≤p+1 n!

The WZ action (6.135) suggests that the tachyon potential is

V (T, T ) ∼ e−T T . (6.136)

We now turn to the non-topological part of the branes-antibranes action, which we will
denote by DBI. We expect to get the same tachyon potential in the DBI part. We now make
the assumption that we can write it via the supercurvature. Since the superconnection and

24
supercurvature appear as part of the structure of the system via the Chan-Paton factors one
may expect this to be the case. However, it is also possible that only the topological part of the
branes-antibranes action can be written using the supercurvature. This is related to the question
whether the superbundle structure is indeed a structure of the brane-antibrane system or only
of its topological part. We will continue with the assumption, bearing in mind that we do not
have a proof for it.
The requirement of being able to write the DBI part using the supercurvature, together with
the requirement of getting the same tachyon potential in the DBI part, uniquely fixes the DBI
action to Z
SDBI = −τ0 dp+1 xTr(treF ) . (6.137)
Tp
τ0 is a normalisation constant given by 2[(p+1)/2] = τ0 emm .
2
The order F of (6.137) is precisely (6.128). Using the form of the curvature (6.134) we have
Tp Z  
SDBI = − d(p+1) x e−T T Tr tr eF . (6.138)
2[(p+2)/2]
Thus, the proposed effective action of the branes-antibranes system, written in terms of the
supercurvature (6.134,) is
S = SDBI + SW Z (6.139)
with SDBI given by (6.138) and SW Z by (6.135).
One can rewrite the supercurvature using Clifford algebra as
 1 µν
iγ µ ∂µ T
  
γ Fµν 1 0
2 2
F =  − (T − m2 ) 
  
 
1 µν
iγ µ ∂µ T 2
γ Fµν 0 1
2 2
= F − (T − m )11 , (6.140)

where we added a constant part.

Consider the WZ part of the non-BPS Dp-brane action. The above discussion of the D-branes
charges associated with the non-BPS Dp-brane imply that the WZ-term is
 
n
−iµp Z p+1 −T 2 X F
SW Z = √ d xe C ∧ Trσ   . (6.141)
2 n≤p+1 n!

That is, the Chern characters of the superbundle encode the Dp-brane charges of condensates
2
on the non-BPS branes. The WZ action suggests that the tachyon potential is V (T ) ∼ e−T .
Following the same reasoning as for the (Dp, Dp) brane system we propose a form of the DBI
part of the non-BPS Dp-brane action, written in terms of the supercurvature as
Tp Z
2
 
SDBI = − √ [(p+2)/2] d(p+1) x e−T Tr tr eF , (6.142)
22
√
with F given by (6.140).
√ The factor 1/ 2 in (6.142) is due to the fact that the tension of a
non-BPS Dp-brane is 2Tp , and the matrix structure of F.
To summarize, the proposed effective action of the non-BPS Dp-brane is S = SDBI + SW Z ,
with SDBI given by (6.142) and SW Z by (6.141).
Consider a non-BPS Dp-brane that carries a D(p-1)-brane charge. Upon tachyon condensation
we expect to get a BPS D(p-1)-brane. The tachyon T is a function of one coordinate transverse
to the expected D(p-1)-brane world volume. Denote this coordinate by x1 = x. We take the
tachyon configuration to be T = αx where α is constant. For such a configuration higher than
two derivatives of the tachyon vanish and we do not have to worry about not including them in
the effective action.

25
 Using the WZ action (6.141) we get the coupling of RR p-form to the Non-BPS-brane. It
reads
√ Z
2
SW Z = 2µp dp+1 x Cp ∂T e−T
Z
= µp−1 dp xCp , (6.143)

with µp−1 = µp 2πls , and we have rescaled to restore the appropriate dimensions. We see that we
get the charge corresponding to the D(p-1)-brane, independently of the form of the gauge field
strength.
Plugging the α → ∞ solution into the effective action, we get
√ Z
p+1 −α2 x2
√ √ √ Z p
Skink = − 2Tp d xe 2
1 + α |α→∞ = (− 2Tp ) π( d y) . (6.144)
√
Therefore the tension of the kink is Tkink = 2π Tp . After restoring the appropriate units we
have √
Tkink = (2π α0 ) Tp ≡ Tp−1 , (6.145)
as the exact value.
As we noted, finite changes of the value of F do not affect the tension of the kink, but infinite
changes will. All the other configurations will not satisfy the BPS relation between the charge
and the tension.
It is worth exploring the kink profile in another set of variables. Consider the DBI action for
the non-BPS Dp-brane proposed
Z q
S = −Tp dp+1 xV (Te ) −det(ηµν + Feµν + ∂µ Te ∂ν Te ) . (6.146)

One can study the kink solutions via this action.

Consider tachyon condensation on a (Dp, Dp) brane system carrying a D(p-2)-brane charge.
The tachyon should form a vortex-like configuration, with the topological charge of the vortex
encoding the D(p − 2) brane charge.
We take the tachyon configuration T = αz, T = αz, where z = x1 + ix2 . Inserting into the
WZ action (6.135) we get the coupling of RR p-form to the BPS-brane. It reads

(2)
Z
1 µ0 ,...µp−1 αβ  
SW Z = µp dp+1 x Cµ0 ...µp−1 (F 1 − F 2 )αβ + 2Dα T Dβ T e−T T
2 p!
Z
1
= µp (2π)(1 + ∆F ) dp−1 x µ0 ...µp−1 Cµ0 ...µp−1 , (6.147)
p!

where ∆F = F 1 − F 2 . Reinstalling 2πα0 one thus finds µcond = 2π µp−2 (1 + ∆F ). Assume that
i
only F12 , i = 1, 2 is different from zero. In order to find the exact charge the vortex-like solution
1 2
should have F12 − F12 = 0.
The correct scaling for the field strength F can be found from calculating the tension of the
vortex. Plugging the α → ∞ solution into the action, we get
√
Z
p+1 −|α|2 |z|2
q
2
1 + β 2 Z p−1
S|vortex = −2Tp d xe 1 + β = −2πTp d x. (6.148)
|α|2

Scaling F such that |β| → |α|2 the tension of the vortex is Tp−2,cond = 2π Tp−2 . After reinstalling
2πα0 one finds
Tp−2 = (2π)2 α0 Tp−2 , (6.149)
which is the correct value of the D(p-2)-brane tension.

26
 Consider, for instance, non-BPS Dp-brane case where the tachyon is real T = T . Up to two
derivatives the action has the structure familiar from BSFT and σ-model perturbation theory
Z
2 1
S = −2Tp d(p+1) xe−T (1 + ∂µ T ∂ µ T ) . (6.150)
2
However, it is easy to see that the numerical coefficients of the higher derivative terms do not
match those of the BSFT action. Indeed, one studies the tachyon condensation with only the
tachyon field excited and one gets the precise tension of the lower-dimensional D(p-1)-brane. In
our variables, we needed a nonzero configuration of the gauge field strength in order to derive
the precise tension of D(p-1)-brane from the kink solution. Upon addition of the gauge fields in
the BSFT formalism there is still a difference between the effective actions.
One can also set T = T = A2 = 0 in the action (6.138), which leads to

Tp Z
1 µν F
S=− d(p+1) x tr e 2 γ µν
, (6.151)
2(p+1)/2
which one can map to the DBI action
Z q
p+1
S = −Tp d x −det(ηµν + Feµν ) , (6.152)

by relating Fe to a formal expansion of sinh(Fµν ). However, as above, one would expect a change
of variables to include the higher derivative terms which were neglected in the slowly varying
fields approximation.
We shall generalize the construction to a vortex solution on a brane-antibrane system. For
this we need to begin with a tachyon effective action on a brane-antibrane system pair. In this
case we have a complex tachyon field T , besides the massless gauge fields A(1) (2)
µ , Aµ and scalar
I I
fields Y(1) , Y(2) corresponding to the transverse coordinates of individual branes. We shall work
with the following effective action that generalizes (5.89):
There have been various other proposals for the tachyon effective action and / or vortex
solutions on brane-antibrane system. The action is
Z q q 
S=− dp+1 x V (T, Y(1)
I I
− Y(2) ) −detA(1) + −detA(2) , (6.153)

where
1 1
(i)
A(i)µν = ηµν + Fµν I
+ φµ Y(i) I
φν Y(i) + (Dµ T )∗ (Dν T ) + (Dν T )∗ (Dµ T ) , (6.154)
2 2
(i)
Fµν = φµ A(i) (i)
ν − φν Aµ , Dµ T = (φµ − iA(1) (2)
µ + iAµ )T , (6.155)
I I 2
and the potential V (T ) depends on |T | and − Y(2)
P
I (Y(1) ) only. For small T , V behaves as

X Y I − Y I 2
  !  
I I 1 (1) (2) 1
V (T, Y(1) − Y(2) ) = Tp 1 + − |T |2 + O(|T |4 ) . (6.156)
2  I 2π 2

Tp denotes the tension of the individual D-p-branes. Although this action has not been derived
from first principles, we note that this obeys the following consistency conditions:

1. The action has the required invariance under the gauge transformation:

T → e2iα(x) T, A(1) (1)

µ → Aµ + φµ α(x), A(2) (2)
µ → Aµ − φµ α(x) . (6.157)

2. For T = 0 the action reduces to the sum of the usual DBI action on the individual branes.

27
 3. If we require the fields to be invaiant under the symmetry (−1)FL that exchanges the brane
and the antibrane, we get the restriction:

T = real, A(1) (2)

µ = Aµ ≡ Aµ ,
I
Y(1) I
= Y(2) ≡YI. (6.158)

Under this restriction the action becomes proportionl to that on a non-BPS D-p-brane, as
given in (5.89). This is a necessary consistency check, as modding out a brane-antibrane
configuration by (−1)FL is supposed to produce a non-BPS D-p-brane[39].

We should keep in mind however that these constraints do not fix the form of the action uniquely.
Nevertheless we shall make the specific choice given in (6.153) and proceed to study the vortex
solution in this theory.
We expect our analysis to be valid for a more general action of the form:
Z hq
(1) (1)
S = − d p+1
xV I
(T, Y(1) − I
Y(2) ) −det(gµν + Fµν )F (Gµν ∗
(1) Dµ T Dν T )
q i
(2) (2)
+ −det(gµν + Fµν )F (Gµν ∗
(2) Dµ T Dν T ) (6.159)

(i)
where gµν I
= ηµν + φµ Y(i) I
φν Y(i) is the induced closed string metric on the ith brane, Gµν
(i) is the
1/2
open string metric on the ith brane and the function F (u) grows as u for large u.
The energy momentum tensor T µν associated with this action is given by:
hq q i
µν µν
T µν = −V (T, Y(1)
I I
− Y(2) ) −det(A(1) )(A−1
(1) )S + −det(A(2) )(A−1
(2) )S . (6.160)

Another surprising feature of both the kink and the vortex solutions is that the world-volume
theory on the soliton has exactly the DBI form without any higher derivative corrections. This
means that all such corrections must come from higher derivative corrections to the original
actions (5.89) and (6.153). This may seem accidental, but may be significant for the following
reasons. This result suggests that there is a close relation between the systematic derivative
(of field strength) expansion of the world-volume action of the non-BPS D-p-brane (D-p-brane -
D-p-brane pair) and that of the BPS soliton solution representing D-(p − 1) brane (D-(p − 2)-
brane). It will be interesting to explore this line of thought to see if one can establish a precise
connection between the two. Since the derivative expansion on the world-volume of BPS D-
branes is well understood, finding a connection of the type mentioned above will provide a
better understanding of the derivative expansion of the world-volume action of a non-BPS D-
brane / brane-antibrane system. We considered DBI-type effective action of a complex tachyon
and gauge fields of U(1)×U(1) symmetry, describing brane-antibrane system with fundamental
strings. In the coincidence limit of (D2, D2), static vortex solutions are obtained. Without
DBI electromagnetic field, there exist only singular static global and local D-vortex solutions.
When the radial component of electric field is turned on, we found regular static global and local
D-vortex solutions. The obtained point-like D-vortex configurations are naturally embedded
in straight stringy solutions in (D3, D3) brane system, and are identified with D-strings (D1-
branes). If the obtained macroscopic D-strings are gravitating, they become naturally candidates
of cosmic D-strings in the early Universe.

7 The (Dp, Dp) Brane System in Dual Effective Actions

Although the complete effective action describing the (Dp, Dp) brane-antibrane system has not
been derived from first principles, it is known to satisfy a set of consistency conditions. In the
context of our discussion in these proceedings this action describes the Higgs phase for the relative
BI vector field. We will work to second order in α0 , ignoring tachyonic couplings to Cp−1 and
taking the other RR potentials to zero. This truncated action will contain however the relevant

28
couplings for describing the most important aspects of the dynamics of the (Dp, Dp) system,
both in the Higgs and in the confining phases. Our starting point is then the action
Z n 1  1
 1
S(χ, A) = dp+1 x e−φ F + + B2 + e−φ F − ∧ ∗F − +
F + + B2 ∧ ∗
2 2 4 o
+|T | (dχ − A ) ∧ ∗(dχ − A ) + d|T | ∧ ∗d|T | − V (|T |) + Cp−1 ∧ F − .
2 − −
(7.161)

Here we have set 2πα0 = 1, A+ and A− are the overall and relative BI vector fields: A+ = A + A0 ,
A− = A − A0 , and the complex tachyon is parametrized as T = |T |eiχ . V (|T |) is the tachyon
potential, whose precise form will be irrelevant for our analysis. The pullbacks of the spacetime
fields into the worldvolume are implicit.
The last coupling shows that when the tachyon condenses in a vortex-like configuration a
D(p − 2)-brane is generated as a topological soliton. In this process the relative U (1) vector
superfield eats the scalar superfield χ, gets a mass and is removed from the low energy spectrum.
The overall U (1) vector superfield, under which the tachyon is neutral, remains unbroken, but it
is believed to be confined.
Note that since A− is massive it cannot be dualized in the standard way. We can however use
the standard procedure to dualize the phase of the tachyon and A+ . These fields are dualized,
respectively, into a (p − 1)-form, Wp−1 , and a (p − 2)-form, that we denote by A− p−2 given that
due to the opposite orientation of the antibrane the relative and overall gauge potentials should
be interchanged under duality. The intermediate dual action that is obtained after these two
dualizations are carried out is such that, up to a total derivative term, A− becomes massless and
can therefore be dualized in the standard way into A+ p−2 .
The final dual action reads:
Z n 1  1 
S(Wp−1 , Ap−2 ) = dp+1 x eφ +
Fp−1 + Wp−1 + Cp−1 ∧ ∗ +
Fp−1 + Wp−1 + Cp−1
2 2
1 φ − − 1 −
o
+ e Fp−1 ∧ ∗Fp−1 + dWp−1 ∧ ∗dWp−1 + d|T | ∧ ∗d|T | − V (|T |) − B2 ∧ F p−1 (7.162)
4 4|T |2

The action (7.162) is an extension of the actions proposed, and it will become clear later
that it describes the confining phase for the overall (p − 2)-form dual potential. This phase
arises after the condensation of zero-dimensional topological defects which originate from the
end-points of open strings stretched between the branes. The interpretation of the low energy
mode Wp−1 is as describing the fluctuations of these defects, and is such that away from the
defects Wp−1 = dA+ p−2 . It can be seen that the original gauge invariance has been mapped into a
gauge transformation of Wp−1 and A+ p−2 . This symmetry can be gauge fixed by absorbing Fp−1
+

into Wp−1 , which becomes then massive. The overall A+ p−2 gauge potential is then removed from
the low energy spectrum through the so-called Julia-Toulouse mechanism, which we will discuss
further in the next section and is, essentially, the contrary of the more familiar Higgs mechanism.
The Julia-Toulouse mechanism is therefore the responsible for the removal of the relative U(1)
at strong coupling. However it clearly sheds no light on the removal of A+ .
When comparing the action (7.162) to the actions describing the confining phases of anti-
symmetric field theories presented one sees that the modulus of the tachyon plays the role of
the density of condensing topological defects, as can be expected since the instability in the con-
fining phase is originated by the presence of the topological defects. In the confining models of
Quevedo and Trugenberger a consistency requirement is that the antisymmetric field theory in
the Coulomb phase is recovered for zero density of defects. This is indeed satisfied by our action
(7.162) for vanishing tachyon, since the |T | → 0 limit forces the condition that Wp−1 must be
exact and can therefore be absorbed through a redefinition of A+ , recovering the Coulomb phase
in dual variables. Quevedo and Trugenberger made explicit in the framework of antisymmetric
field theories an old idea in solid-state physics due to Julia and Toulouse. These authors argued
that for a compact tensor field of rank (h − 1) in (p + 1)-dimensions a confined phase might

29
arise after the condensation of (p − h − 1)-dimensional topological defects. The fluctuations of
the continuous distribution of topological defects generate a new low-energy mode in the theory
which can be described by a new h-form, Wh , such that away from the defects Wh = dAh−1 ,
where Ah−1 is the original tensor field. The effective action describing the confining phase of the
antisymmetric tensor field then depends on a gauge invariant combination of the antisymmetric
tensor field, Ah−1 , and the extended h-form, Wh . This combination is such that when the den-
sity of topological defects vanishes the original action describing the antisymmetric tensor field
theory in the Coulomb phase is recovered. As discussed, the finite condensate phase is a natural
generalization of the confinement phase for a four dimensional vector gauge field to arbitrary
(h − 1)-forms in d dimensions.
Given that the worldvolume theory of a (Dp, Dp) system is a vector field theory, the results
for h = 2 can be applied to this case. In this case the Coulomb phase is the phase with zero
tachyon, and it is therefore described by the Lagrangian:
1  1  1
L(A) = e−φ F + + B2 ∧ ∗ F + + B2 + e−φ F − ∧ ∗F − + Cp−1 ∧ F − . (7.163)
2 2 4
Developping now on the ideas in this rseaerch article we have that the topological defects
whose condensation will give rise to the confining phase are (p − 3)-branes, which originate in
this case from the end-points of D(p − 2)-branes stretched between the Dp and the Dp. The new
mode associated to the fluctuations of the defects is described by a 2-form, W2 , which will couple
in the action through a gauge invariant combination with the overall U (1) vector superfield. The
action should depend as well on the density of topological defects, such that when this density
vanishes the original action in the Coulomb phase, given by (7.163), is recovered. We will see
that, contrary to the actions constructed, where the density of topological defects entered as
a parameter which was interpreted as a new scale in the theory, in the (Dp, Dp) case it must
be a dynamical quantity because it is related through duality to the modulus of the tachyonic
excitation of the open D(p − 2)-branes in the dual Higgs phase. We will denote this field by |Te |
and, moreover, we will use the duality with the Higgs phase to include in the action its kinetic
and potential terms.
The dual effective action that we propose for describing the confining phase of the (Dp, Dp)
brane system is then given by
Z n 1  1  1
S(W2 , A) = dp+1 x e−φ F + + W2 + B2 ∧ ∗ F + + W2 + B2 + e−φ F − ∧ ∗F − +
2 2 4
1 o
+ e dW2 ∧ ∗dW2 + d|Te | ∧ ∗d|Te | − V (|Te |) + Cp−1 ∧ F − . (7.164)
4|T |2
This action has been constructed under four requirements. One is gauge invariance, both under
gauge transformations of the BI vector fields and under W2 → W2 + dΛ1 , which ensures that only
the gauge invariant part of W2 describes a new physical degree of freedom. This transformation
must be supplemented by A+ → A+ − 2Λ1 , a symmetry that has to be gauge fixed. The second is
relativistic invariance. The third requirement is that the original action describing the Coulomb
phase must be recovered when |Te | → 0. Indeed, when |Te | → 0 we must have that dW2 = 0, so
that W2 = dψ1 for some 1-form ψ1 . This form can then be absorbed by A+ , and the original action
(7.163) is recovered. On the other hand, consistency with the duality symmetries of superstring
theory will later on imply that W2 must couple only to the overall U (1) vector field.
Now, in (7.164) F + can be absorbed by W2 , fixing the gauge symmetry, and the action can
then be entirely formulated in terms of W2 and the relative vector field:

−
Z n     1
S(W2 , A ) = dp+1 x e−φ W2 + B2 ∧ ∗ W2 + B2 + e−φ F − ∧ ∗F − +
4
1 o
+ e dW2 ∧ ∗dW2 + d|Te | ∧ ∗d|Te | − V (|Te |) + Cp−1 ∧ F − . (7.165)
4|T |2

30
In this process the original gauge field A+ has been eaten by the new gauge field W2 , and has
therefore been removed from the low energy spectrum. This solves the puzzle of the unbroken
overall U (1) at weak string coupling through the Julia-Toulouse mechanism. Let us now see how
the fundamental superstring arises from this dual effective action.
Inspired by Mandelstam-’t Hooft duality we expect that the dual of the action (7.164) de-
scribes the Higgs phase for the (p − 2)-form field dual to the overall BI vector. The dualization of
the BI vector fields in (7.164) takes place in the standard way, given that they only couple through
their derivatives. In turn, the 2-form W2 is massive, but it can still be dualized in the standard
way from the intermediate dual action that is obtained after dualizing the BI vector fields, in
which it only couples through its derivatives. Let us call the dual of this form, a (p − 3)-form,
χp−3 . The final dual action reads:
Z n 1  1  1 − −
S(χp−3 , Ap−2 ) = dp+1 x eφ +
Fp−1 + Cp−1 ∧ ∗ +
Fp−1 + Cp−1 + eφ Fp−1 ∧ ∗Fp−1
   2  2 4 o
+|Te |2 dχp−3 − A−
p−2 ∧ ∗ dχ p−3 − A−
p−2 + d|Te | ∧ ∗d|Te | − V (|Te |) − B ∧ F −
2 p−1 (7.166)
where once again the overall and the relative gauge superfields are interchanged.
The action (7.166) describes an Abelian Higgs model for the relative (p − 2)-form field, with
the dual (p − 3)-form χp−3 playing the role of the associated Goldstone boson. That an effective
mass gauge invariant term of this kind could drive the dual Higgs mechanism was suggested,
although it could not be explicitly derived from the action describing the Higgs phase at weak
coupling, i.e. from Sen’s action. The Goldstone boson χp−3 is associated to the fluctuations of the
(p − 3)-dimensional topological defects that originate from the end-points of the D(p − 2)-branes
stretched between the Dp and the Dp. This is consistent with the fact that this field is the
worldvolume dual of the field W2 , which was accounting for these fluctuations in the confining
action (7.164). Moreover, we can identify for p = 3 the condensing Higgs scalar as the modulus of
the tachyonic mode associated to open D-strings stretched between the D3 and the D3. Indeed
when p = 3 the action (7.166) turns out to be the S-dual of the original action (7.161) describing
the perturbative Higgs phase of the (D3, D3) system. This is an important consistency check,
although strictly speaking S-duality invariance would only be expected for zero tachyon. In this
duality relation the modulus of the perturbative tachyon is mapped into |Te |, which can then be
interpreted as the modulus of the tachyonic excitation associated to the open D-strings. Since
χe has also an interpretation as the phase of the dual tachyon we can think of Te as the complex
tachyonic mode associated to the D-strings stretched between the D3 and the D3. For p 6= 3,
since the tachyonic condensing charged object is a (p − 3)-brane, the phase of the tachyon is
replaced by a (p − 3)-form. It would be interesting to clarify the precise way in which these fields
arise as open D(p − 2)-brane modes.
As we have seen, a (Dp, Dp) system admits two types of topological defects: particles and
(p − 3)-branes, which are, respectively, perturbative and non-perturbative in origin. The com-
bined electric and magnetic Higgs mechanisms introduce mass gaps to both U(1) vector poten-
tials, being the only remnants D(p − 2)-branes and fundamental strings, realized as solitons on
the common (p + 1)-dimensional worldvolume. We have seen that it is possible to incorporate the
non-perturbative degrees of freedom associated to the extended topological defects in the weak
coupling regime, using Julia and Toulouse’s idea, introducing a new form describing the fluctu-
ations of these defects and imposing a set of consistency conditions. In fact, one can combine
the weakly coupled action presented in section 3 with Sen’s action in order to incorporate the
degrees of freedom associated to both the zero dimensional and extended topological defects:
Z n 1  1  1
S(χ, W2 , A) = dp+1 x e−φ F + + W2 + B2 ∧ ∗ F + + W2 + B2 + e−φ F − ∧ ∗F − +
2 2 4
1
+|T |2 (dχ − A− ) ∧ ∗(dχ − A− ) + d|T | ∧ ∗d|T | + e dW2 ∧ ∗dW2 +
4|T |2
o
+d|Te | ∧ ∗d|Te | − V (|T |) − V (|Te |) + Cp−1 ∧ F − . (7.167)

31
This action describes both the perturbative and the non-perturbative Higgs mechanisms simul-
taneously at weak coupling, and it admits both a magnetic vortex solution, which by charge
conservation is identified with the D(p − 2)-brane, and an electric vortex solution, identified as
the fundamental string.
The total higher-dimensional dual effective action of extremal (Dp, Dp) brane systems is
constructed in the special form
Z n 1  1  1
ST (χ, A+ +
− , W− ) = dp+1 x e−φ F + + B2 ∧ ∗ F + + B2 + e−φ F − ∧ ∗F −
2 2 4 o
+ |T | (dχ − A ) ∧ ∗(dχ − A ) + d|T | ∧ ∗d|T | − V (|T |) + Cp−1 ∧ F −
2 − −

n 1  1 1 
− −
+ eφ Fp−1 +
+ Wp−1 + Cp−1 ∧ ∗ Fp−1 +
+ Wp−1 + Cp−1 + eφ Fp−1 ∧ ∗Fp−1
2 2 4
1 −
o
+ dW p−1 ∧ ∗dW p−1 + d|T | ∧ ∗d|T | − V (|T |) − B 2 ∧ F p−1
4|T |2
n 1  1  1
+ e−φ F + + W2 + B2 ∧ ∗ F + + W2 + B2 + e−φ F − ∧ ∗F −
2 2 4
1 −
o
+ dW 2 ∧ ∗dW 2 + d|Te | ∧ ∗d|Te | − V (|Te |) + C
p−1 ∧ F
4|Te |2
n     1
+ e−φ W2 + B2 ∧ ∗ W2 + B2 + e−φ F − ∧ ∗F −
4
1 −
o
+ dW 2 ∧ ∗dW 2 + d|Te | ∧ ∗d|Te | − V (|Te |) + C
p−1 ∧ F
4|Te |2
n 1  1  1 − −
+ eφ Fp−1 +
+ Cp−1 ∧ ∗ Fp−1 +
+ Cp−1 + eφ Fp−1 ∧ ∗Fp−1
2   2  4 o
+ |Te |2 dχp−3 − A− p−2 ∧ ∗ dχ p−3 − A −
p−2 + d| e | ∧ ∗d|Te | − V (|Te |)
T
n 1  1  1
−
− B2 ∧ Fp−1 + e−φ F + + W2 + B2 ∧ ∗ F + + W2 + B2 + e−φ F − ∧ ∗F −
2 2 4
2 − − 1
+ |T | (dχ − A ) ∧ ∗(dχ − A ) + d|T | ∧ ∗d|T | + e dW2 ∧ ∗dW2
4|T |2
o
+ d|Te | ∧ ∗d|Te | − V (|T |) − V (|Te |) + Cp−1 ∧ F − . (7.168)

8 Conclusion
In this paper we have analyzed kink and vortex solutions in tachyon effective field theory by
postulating suitable form of the tachyon effective action on the non-BPS D-brane and brane-
antibrane system respectively. In both cases the topological soliton has all the right properties
for describing a BPS D-brane. These properties include localization of the energy-momentum
tensor on subspaces of the codimensions, as is expected of a D-brane and also the DBI form of
the effective action describing the world-volume theory on the soliton. For the kink solution we
have also done the analysis including the world-volume fermions, and shown the appearance of
κ-symmetry in the world-volume theory on the kink. One feature of both the solutions is infinite
spatial gradient of the tachyon field away from the core of the soliton. If we want to construct a
solution describing tachyon matter in the presence of such a soliton, then the spatial gradient of
the tachyon field represents local velocity of the tachyon matter. This shows that tachyon matter
in the presence of such a solution will fall towards the core of the soliton. If this feature survives
in the full superstring theory, then it will imply that any tachyon matter in contact with the
soliton will be sucked in immediately. This is consistent with the general analysis where similar
effect was found by analyzing the boundary state associated with the time dependent solutions.
This might provide a very effective means of absorbing tachyon matter from the surrounding by
a defect brane, and drastically modify the results for the formation of topological defects during

32
the rolling of the tachyon field. The appearance of infinite slope during the dynamical process of
defect formation has already been observed in present scientific research. We should note how-
ever that a different type of solution where a codimension soliton and tachyon matter coexist has
been constructed. One question that we have not addressed is the analysis of the world-volume
theories on multiple kink-antikink pairs and multivortex solutions. In a finite region around the
location of each soliton the solution will have the form discussed, and we need to ensure that be-
fore taking the limit, the various fields match smoothly, keeping or order larger in the intervening
space. Analysis of the world-volume theory around such a background will clearly yield the sum
of the world-volume actions on the individual solitons, since essentially the field configurations
around individual solitons do not talk to each other in the limit. The interesting question is
whether we can see the excitations associated with the fundamental string stretched between the
solitons. We believe these excitations must come from classical solutions describing fundamental
superstring along the line. We can, for example, take the solutions in the DBI theory given in
and lift them to solutions of the equations of motion derived from the current considerations.
The spontaneously broken gauge symmetry that mixes the states of the open superstring living
on individual D-branes with states of the open superstring stretched between different D-branes,
exchanges perturbative states with solitonic states, and hence is analogous to the electric mag-
netic duality symmetry in gauge field theories. Tachyon potential with such highly restricted
behavior has not been available when looked for in the context of the open superstring field
theory constructions. In this sense, it is quite promising that the supergravity analogue of the
tachyon potential obtained in the present work can serve as a successful candidate for such a
flat potential since, as we have seen in this work, it possesses purely logarithmic dependence
on the tachyon field and hence is slowly varying. Thus it will be the issue of one of our future
works to explore the potentially successful role played by the tachyon field arising in the unstable
brane-antibrane system and having the structure of potential discovered in the present work. To
summarize, in this work, using an exact supergravity solution representing the brane-antibrane
system, we demonstrated in a rigorous fashion that one can construct a supergravity analogue of
the tachyon potential which may possess generic features of the genuine stringy tachyon poten-
tials. In doing so, our philisophy was to evaluate the interaction energy between the brane and
the antibrane for small but finite inter-brane separation and then identify it with the tachyon
potential. This identification demands an appropriate suggestion to relate the parameter in the
supergravity solution representing the inter-brane distance to the tachyon field expectation value
and we proposed an ansatz given in the article. The general lesson that one could learn from
the results of this paper is that for many purposes, it is useful to complement the supergravity
action, describing low energy effective action of closed superstring theory, by coupling it to the
tachyon effective action of the type described in this paper. In such a theory, BPS D-branes arise
naturally as topological solitons rather than having to be added by hand, and we get the correct
low energy effective action on these D-branes. Furthermore, we have seen earlier that this effec-
tive action is capable of describing certain time dependent solutions of open string theory, and
solutions describing the fundamental string. Coupling the tachyon field to supergravity does not
give rise to any new perturbative physical states, and hence does not violate any known result in
string theory. Finally, as was argued in the research literature, coupling of the tachyon effective
action to gravity may resolve some of the conceptual problems involving ‘time’ in quantum grav-
ity. Brane-antibrane systems hold the divine key and open the magical gate of knowledge to the
unraveling and direct entry of mankind into the extreme hyperspace where the branes live, swim
and interact under extreme conditions in higher-dimensional moduli painting supercollection with
hypermanifolds of membrane universes. The considered fundamental membrane models provide
the unique opportunity for in-depth construction of multiversum doctrina dominum in the mirror
of a fundamental membrane theory possessing the qualities to accomplished, summarizing and
unifying the general theoretical framework with the beam penetrating deep into the depths of
the magical world structure plus vigorously extremal multiverse.

33
References
[1] A. Sen, “Supersymmetric world-volume action for non-BPS D-branes,” JHEP 9910, 008
(1999) [arXiv:hep-th/9909062].

[2] E. A. Bergshoeff, M. de Roo, T. C. de Wit, E. Eyras and S. Panda, “T-duality and actions
for non-BPS D-branes,” JHEP 0005, 009 (2000) [arXiv:hep-th/0003221].

[3] J. Kluson, “Proposal for non-BPS D-brane action,” Phys. Rev. D 62, 126003 (2000)
[arXiv:hep-th/0004106].

[4] J. A. Minahan and B. Zwiebach, “Gauge fields and fermions in tachyon effective field theo-
ries,” JHEP 0102, 034 (2001) [arXiv:hep-th/0011226].

[5] M. Alishahiha, H. Ita and Y. Oz, “On superconnections and the tachyon effective action,”
Phys. Lett. B 503 (2001) 181 [arXiv:hep-th/0012222].

[6] N. D. Lambert and I. Sachs, “Tachyon dynamics and the effective action approximation,”
Phys. Rev. D 67, 026005 (2003) [arXiv:hep-th/0208217].

[7] V. Schomerus, “D-branes and deformation quantization,” JHEP 9906, 030 (1999)
[arXiv:hep-th/9903205].

[8] S. Sugimoto and S. Terashima, “Tachyon matter in boundary string field theory,” JHEP
0207, 025 (2002) [arXiv:hep-th/0205085].

[9] J. A. Minahan, “Rolling the tachyon in super BSFT,” JHEP 0207, 030 (2002) [arXiv:hep-
th/0205098].

[10] D. Kutasov, M. Marino and G. W. Moore, “Remarks on tachyon condensation in superstring

field theory,” arXiv:hep-th/0010108.

[11] D. Ghoshal, “Normalization of the boundary superstring field theory action,” arXiv:hep-
th/0106231.

[12] A. Sen, “SO(32) spinors of type I and other solitons on brane-antibrane pair,” JHEP 9809,
023 (1998) [arXiv:hep-th/9808141].

[13] A. Sen, “BPS D-branes on non-supersymmetric cycles,” JHEP 9812, 021 (1998) [arXiv:hep-
th/9812031].

[14] P. Horava, “Type IIA D-branes, K-theory, and matrix theory,” Adv. Theor. Math. Phys. 2,
1373 (1999) hep-th/9812135.

[15] J. M. Cline and H. Firouzjahi, “Real-time D-brane condensation,” arXiv:hep-th/0301101.

[16] M. Cederwall, A. von Gussich, B.E. Nilsson and A. Westerberg, “The Dirichlet super-
three-brane in ten-dimensional type IIB supergravity,” Nucl. Phys. B490, 163 (1997) hep-
th/9610148.

[17] M. Aganagic, C. Popescu and J.H. Schwarz, “D-brane actions with local kappa symmetry,”
Phys. Lett. B393, 311 (1997) hep-th/9610249.

[18] M. Cederwall, A. von Gussich, B.E. Nilsson, P. Sundell and A. Westerberg, “The Dirichlet
super-p-branes in ten-dimensional Nucl. Phys. B490, 179 (1997) hep-th/9611159.

[19] E. Bergshoeff and P.K. Townsend, “Super D-branes,” Nucl. Phys. B490, 145 (1997) hep-
th/9611173.

34
[20] M. Aganagic, C. Popescu and J.H. Schwarz, “Gauge-invariant and gauge-fixed D-brane
actions,” Nucl. Phys. B495, 99 (1997) hep-th/9612080.

[21] N. D. Lambert and I. Sachs, “On higher derivative terms in tachyon effective actions,” JHEP
0106, 060 (2001) [arXiv:hep-th/0104218].

[22] T. Asakawa, S. Sugimoto and S. Terashima, “Exact description of D-branes via tachyon
condensation,” JHEP 0302, 011 (2003) [arXiv:hep-th/0212188].

[23] K. Hashimoto and S. Hirano, “Metamorphosis of tachyon profile in unstable D9-branes,”

Phys. Rev. D 65, 026006 (2002) [arXiv:hep-th/0102174].

[24] G. Arutyunov, S. Frolov, S. Theisen and A. A. Tseytlin, “Tachyon condensation and uni-
versality of DBI action,” JHEP 0102, 002 (2001) [arXiv:hep-th/0012080].

[25] K. Hashimoto and S. Nagaoka, “Realization of brane descent relations in effective theories,”
Phys. Rev. D 66, 026001 (2002) [arXiv:hep-th/0202079].

[26] G. W. Gibbons, K. Hori and P. Yi, “String fluid from unstable D-branes,” Nucl. Phys. B
596, 136 (2001) [arXiv:hep-th/0009061].

[27] P. Mukhopadhyay and A. Sen, “Decay of unstable D-branes with electric field,” JHEP 0211,
047 (2002) [arXiv:hep-th/0208142].

[28] A. Sen, “Some issues in non-commutative tachyon condensation,” JHEP 0011, 035 (2000)
[arXiv:hep-th/0009038].

[29] A. Sen, “Fundamental strings in open string theory at the tachyonic vacuum,” J. Math.
Phys. 42, 2844 (2001) [arXiv:hep-th/0010240].

[30] G. Gibbons, K. Hashimoto and P. Yi, “Tachyon condensates, Carrollian contraction of

Lorentz group, and fundamental strings,” JHEP 0209, 061 (2002) [arXiv:hep-th/0209034].

[31] O. Bergman, K. Hori and P. Yi, “Confinement on the brane,” Nucl. Phys. B 580, 289 (2000)
[arXiv:hep-th/0002223].

[32] J. A. Minahan and B. Zwiebach, “Effective tachyon dynamics in superstring theory,” JHEP
0103, 038 (2001) [arXiv:hep-th/0009246].

[33] A. A. Gerasimov and S. L. Shatashvili, “Stringy Higgs mechanism and the fate of open
strings,” JHEP 0101, 019 (2001) [arXiv:hep-th/0011009].

[34] E. Witten, “D-branes and K-theory,” JHEP 12, 019 (1998) hep-th/9810188.

[35] P. Kraus and F. Larsen, “Boundary string field theory of the DD-bar system,” Phys. Rev.
D 63, 106004 (2001) [arXiv:hep-th/0012198].

[36] T. Takayanagi, S. Terashima and T. Uesugi, “Brane-antibrane action from boundary string
field theory,” JHEP 0103, 019 (2001) [arXiv:hep-th/0012210].

[37] N. T. Jones and S. H. Tye, “An improved brane anti-brane action from boundary superstring
field theory and multi-vortex solutions,” JHEP 0301, 012 (2003) [arXiv:hep-th/0211180].

[38] D. P. Jatkar, G. Mandal and S. R. Wadia, “Nielsen-Olesen vortices in noncommutative

Abelian Higgs model,” JHEP 0009, 018 (2000) [arXiv:hep-th/0007078].

[39] A. Sen, “Non-BPS states and branes in string theory,” arXiv:hep-th/9904207.

35
[40] C. Kennedy and A. Wilkins, “Ramond-Ramond couplings on brane-antibrane systems,”
Phys. Lett. B 464, 206 (1999) [arXiv:hep-th/9905195].

[41] A. Sen, “Tachyon matter,” JHEP 0207, 065 (2002) [arXiv:hep-th/0203265].

[42] M. Gutperle and A. Strominger, “Spacelike branes,” JHEP 0204, 018 (2002) [arXiv:hep-
th/0202210].

[43] A. Strominger, “Open string creation by S-branes,” arXiv:hep-th/0209090.

[44] A. Ishida and S. Uehara, “Rolling down to D-brane and tachyon matter,” arXiv:hep-
th/0301179.

[45] P. Yi, “Membranes from five-branes and fundamental strings from Dp branes,” Nucl. Phys.
B 550, 214 (1999) [arXiv:hep-th/9901159].

[46] C. G. Callan and J. M. Maldacena, “Brane dynamics from the Born-Infeld action,” Nucl.
Phys. B 513, 198 (1998) [arXiv:hep-th/9708147].

[47] G. W. Gibbons, “Born-Infeld particles and Dirichlet p-branes,” Nucl. Phys. B 514, 603
(1998) [arXiv:hep-th/9709027].

36