Light-Cone Superstring in Ads Space-Time: R.R. Metsaev, C. B. Thorn and A.A. Tseytlin
Light-Cone Superstring in Ads Space-Time: R.R. Metsaev, C. B. Thorn and A.A. Tseytlin
Light-Cone Superstring in Ads Space-Time: R.R. Metsaev, C. B. Thorn and A.A. Tseytlin
OHSTPY-HEP-T-00-017
UFIFT-HEP-00-26
hep-th/0009171
Abstract
We consider fixing the bosonic light-cone gauge for string in AdS space in the
phase space framework, i.e. by choosing x+ = , and by choosing so that
P + is distributed uniformly (its density P + is independent of ). We discuss
classical bosonic string in AdSd and superstring in AdS5 S 5 . In the latter case
the starting point is the action found in hep-th/0007036 where the -symmetry is
fixed by a fermionic light-cone type gauge. We derive the light-cone Hamiltonian
in the AdS5 S 5 case and in the case of superstring in AdS3 S 3 . We also obtain
a realization of the generators of the basic superalgebra psu(2, 2|4) in terms of
AdS5 S 5 superstring coordinate 2-d fields in the light-cone gauge.
Introduction
To better understand string theory duals of gauge theories with various amounts of
supersymmetry [1] it is important to make progress in what should be the simplest case
(large N) N =4 Super Yang-Mills theory dual to (weakly coupled) type IIB superstring
theory in AdS5 S 5 space with Ramond-Ramond flux [2]. Though this background has
a lot of symmetry, solving the corresponding string theory appears to be a complicated
problem. The commonly used procedure, in the known exactly solvable cases, is to start
with a string action, solve the classical string equations, then quantize the theory, find
the string spectrum, vertex operators, scattering amplitudes, etc. Even the first steps in
this program are nontrivial in the AdS5 S 5 string case.
To start with, the classical equations of bosonic string in AdS space, while completely integrable, are not explicitly solvable (in contrast to, e.g., the case of string
on a group manifold described by the WZW model). The presence of a curved R-R
background indicates that one should use the manifestly supersymmetric Green-Schwarz
[3] description of superstring. Finding an explicit expression for the curved space GS
action is difficult in general (one needs to know the component expansion of the background supergravity superfields). In the present AdS5 S 5 case, this technical problem has a nice geometrical solution based on viewing string as moving on the supercoset P SU(2, 2|4)/[SO(1, 4) SO(5)] which replaces the flat superspace (10-d super
Poincare)/SO(1, 9) [4]. The resulting action, though explicitly known [46], is highly
nonlinear containing terms of many orders in . The fermionic part of the action simplifies dramatically in proper -symmetry gauges it becomes quadratic and quartic in
only [79]. Its structure is similar to that of the flat space GS action in a covariant
-symmetry gauge which is also quartic in fermions (see e.g.[10]).
1.1
Review
To illustrate this point, let us consider the 4-d Lorentz covariant S-gauge [11] which
leads to an action equivalent to the one in [7, 8] but uses the fermionic parametrization
adopted in [11] and here, which is useful for comparison with the light-cone gauge actions
below (see Appendix C of [11] for details). If one interprets the AdS5 S 5 supergroup
P SU(2, 2|4) as the N = 4 superconformal group in 4 dimensions, it is natural to split
the fermionic generators into 4 standard supergenerators Qi and 4 special conformal
supergenerators Si (we suppress the 4-d spinor indices). The associated superstring
coordinates will be denoted as i and i . The covariant S-gauge corresponds to setting
all i to zero. The resulting superstring Lagrangian written in the 4+6 parametrization
in which the metric of AdS5 S 5 is ds2 = Y 2 dxa dxa + Y 2 dY M dY M (a = 0, ..., 3; M =
1, ..., 6) has the following simple structure [11]1
L =
1 2 a a
M
gg Y L L + Y 2 Y M Y M (i Y M ai
ij aj + h.c.) , (1.1)
2
The actions in [7, 8] have an isomorphic form, corresponding to a specific choice of the 10-d Dirac
matrix representation.
L=
hi
i
1 A
g x xA
g x+ (i i + i i ) x+ i Cij j + h.c. , (1.2)
2
2
where xA are 10 flat bosonic coordinates and Cij is a constant charge conjugation matrix.
Note that while the s and s enter diagonally in the kinetic term, they are mixed in
the WZ term. This form of the original GS action is the flat space limit of the light-cone
AdS5 S 5 action of [11] (cf. (1.1))
i
h
1
L = gg Y 2 ( x+ x + x x) + Y 2 D Y M D Y M
2
i
h
i 2
gg Y x+ i i + i i + i i + i i + iY 2 x+ ( 2 )2
2
i
h
j
M
j
2|Y
|
x)
+
h.c.
,
+ |Y | x+ i M
Y
(
ij
(1.3)
D Y M Y M + ii (M N )i j j Y N Y 2 x+ .
Here we split the 4-d coordinates as xa = (x+ , x , x, x), and the matrix M is the same
as in (1.1) (M N is the commutator of M and its conjugate, see Appendix A).
2
For a different approach to light-cone type -symmetry gauge fixing in the AdS5 S 5 string action
see [15].
This action contains x only in the bosonic part and only linearly, and the fermionic
kinetic terms are multiplied by the derivative of x+ only. One expects, therefore, that
after one fixes the bosonic light-cone gauge, it should lead to a well defined starting point
for quantizing the theory in the short string sector. As was pointed out in [11, 17, 18]
(see also Section 2), in the case of the AdS type curved spaces, the bosonic light-cone
gauge x+ = in the Polyakov string action can not be combined with the standard
1.2
Summary
The absence of manifest 2-d Lorentz symmetry suggests that it is natural to use the phase
space formulation of the light-cone gauge fixed theory. To develop such a formulation
is the aim of the present paper. We shall explain how the original phase space GGRT
approach [19] to light-cone gauge fixing (x+ = , P + = 1) can be directly applied to the
present case of a curved background.
Most of our discussion will be rather formal, with possible applications depending
on insights into the structure of the resulting AdS5 S 5 string light-cone Hamiltonian.
To summarize, the light-cone phase space Lagrangian, corresponding to (1.3) that we
obtain below, is given by (see (C.5))
i
L = P x + PM Y M + p+ (i i + i i + i i + i i) + P ,
2
(1.4)
where P , PM are the canonical momenta for x = (x, x) and Y M , and the light-cone
Hamiltonian density is
H = P =
i
1 h 2
4
4 2
M Y M +Y 2 (p+2 ( 2 )2 +2ip+ M N YM PN )
P
+|Y
|
P
P
+|Y
|
x
+
Y
M
M
2p+
i
h
j
M j
2|Y
|
x
)
+
h.c.
.
(1.5)
|Y | i M
Y
(
i
ij
As usual for light-cone string, the coordinate x does not appear in the Hamiltonian,
but is determined in terms of the coordinates that do appear via the reparametrization
constraint
i
p+ x + P x + PM Y M + p+ (i i + i i + i i + i i ) = 0 .
(1.6)
2
For a closed piece of string, the integral of this constraint over constrains the state
space to the subspace invariant under translations.
Restoring the dependence on and the scale R of AdS5 S 5 one can then analyze
various limits, e.g., (i) particle theory limit ( 0), (ii) null string limit ( ), 3
(iii) AdS5 S 5 supergravity or strong t Hooft coupling SYM limit ( /R2 0), (vi)
weak t Hooft coupling SYM limit ( /R2 ). In the particle theory limit, the string
3
As in the flat space case [20], the null string limit is obtained by dropping terms with derivatives
with respect to the world sheet coordinate .
Twistors are very helpful in the construction of the theory of interacting massless higher spin fields
in AdS space [22, 23], which has certain similarities with string theory.
of unit 6-d vector) which are useful for translation between the corresponding forms of the
AdS5 S 5 superstring action. In Appendix C we review the two forms of the -symmetry
light-cone gauge fixed action which use the 4+6 parametrizations of AdS5 S 5 [11]
and write down the corresponding phase space Lagrangians. In Appendix D we present
some details of the construction of conformal supercharges in Section 6.
Let us start with a review of some previous discussions of light-cone gauge fixing for
bosonic strings in curved space. In flat space in BDHP formulation [37, 38] one starts by
fixing the conformal gauge
= ,
gg , g detg , det = 1 ,
(2.1)
and then fixes the residual conformal diffeomorphism symmetry on the plane by choosing x+ (, ) = .5 An alternative (equivalent) approach is to use the original GGRT
[19] formulation based on writing the Nambu action in the first order form and fixing
the diffeomorphisms by the two conditions on one coordinate and on one canonical
momentum:
x+ = ,
P + = const .
(2.2)
The obvious requirement for being able to choose the light-cone gauge in a curved space
is the existence of a null Killing vector. The first approach based on fixing the conformal
gauge for the 2-d metric does not in general apply in curved spaces with null Killing
vectors which are not of the direct product form R1,1 M d2 . The exception is the case
when the null Killing vector is covariantly constant [40]. One is thus forced to give up
the standard conformal gauge (2.1) and fix the diffeomorphisms by imposing instead,
e.g., 00 = 1, x+ = . This gauge choice is consistent provided the background metric
satisfies G+ = 1, G = G = 0, Gmn = 0 [17].
The above conditions do not apply in the case of the AdS-type metric (x stands for
all d 3 transverse coordinates):
ds2 = G()dxa dxa + d2 = G()(dt2 + dx2 + dx23 ) + d2 .
(2.3)
Indeed, the two null Killing vectors here are not covariantly constant and also G+ =
G() 6= 1. However, a slight modification of the above gauge conditions of [17] on 00 , x+
represents a consistent gauge choice [11]6
G() 00 = 1 ,
x+ = .
(2.4)
There is a potential complication that the norm of the Killing vector may vanish if there is
a horizon where G = 0. We shall adopt a formal approach, assuming that the degeneracy
5
For a discussion of various ways of fixing the light-cone gauge in the case of flat target space see,
e.g., [39].
6
A closely related light-cone gauge choice = diag(G1 , G), x+ = was originally suggested by
A. Polyakov [18].
Pa = h00 x a h01 xa ,
= h00 G1 h01 G1 .
(2.6)
1
(P2 + 2P + P ) + G2 ()(
x2 + 2
x+ x ) + G()(2 + 2 )
00
2h
h01
+ x+ P + x P + ) ,
(
x P +
h00
(2.7)
where 1/h00 and h01 /h00 play the role of the Lagrange multipliers. As in flat space, x+
is a natural choice for the evolution parameter.9 This suggests choosing the light-cone
gauge as in (2.2)
x+ = ,
7
P + = p+ l1 = const .
(2.8)
The Poincare parametrization was used in the light-cone formulation of field dynamics in AdS
spacetime [41, 42].
8
This subtlety and the issue of fixing a global diffeomorphism gauge for AdS string was discussed in
[44].
9
In [41, 42] it was demonstrated that in field theory in AdS space in Poincare coordinates x+ can be
treated as a light-cone evolution parameter. Therefore, it is reasonable to consider x+ as a light-cone
evolution parameter in AdS string theory too.
For flexibility, we introduced the length l of the string parameter .l may be chosen to
be equal to p+ (which is useful in the study of interactions) or to 2 = 1 (which is
useful in the study of the spectrum). In Section 3 we shall set l = p+ , while in Sections
4-6 we shall use l = 1. To transform the expressions in Sections 4-6 into the l = p+
scheme (P + = 1) one is simply to put p+ = 1 in the Lagrangian and Hamiltonian
densities.
Integrating over P we get the relation
h00 = p+ l1 .
(2.9)
=0,
p+ l1 x + x P +
(2.10)
Note that this is equivalent (for l = p+ ) to the gauge condition on the 2-d metric in
(2.4). The expression for P follows from the 1/h00 constraint. Integration over the
non-constant part of h01 /h00 leads to
=0.
d(
x P + )
(2.11)
2
2
2
2
2
P = +
d P + G ()
x + G()( + ) .
(2.12)
2p 0
R
Note that the equation for x implies that (h01 /h00 ) = 0. Since it also implies h01 = 0
at open string ends, it follows that h01 = 0 in the case of open string. But we can only
conclude that h01 =const in the case of closed string. This constant Lagrange multiplier
imposes the integrated diffeomorphism constraint (2.11), necessary for consistency in the
closed string case.
Let us note that fixing the light-cone gauge in the action, before obtaining the equations of motion, results in lost equations of motion which would be obtained by varying
x+ and P + . It is easy to check that these equations (obtained before gauge fixing)
determine x and P , the time derivatives of variables that have been eliminated by
the diffeomorphism constraints. Indeed, the expressions they give for P and x are
equivalent to ones obtained by differentiating the constraints with respect to time. Thus
the only new information
contained in the lost equations is information about the zero
Rl
mode of x , namely, 0 d x = pl+ P . Since x does not enter the gauge-fixed Hamiltonian this new information is unnecessary for subsequent analysis of the dynamics of the
system.
The bosonic string in AdS space is not consistent (conformally invariant) at the
quantum level. To make it conformally invariant one may add extra couplings to an
NS-NS 3-form background or add fermions and consider superstring coupled to an R-R
background. The simplest example is bosonic string propagating on the SL(2, R) group
manifold described by the WZW model. It is instructive to demonstrate how the phase
space light-cone gauge approach described above applies in this case. In the standard
7
Gauss parametrization, the Lagrangian of the SL(2, R) WZW model is the same as (2.5)
(with G = e2 ) plus a WZ term
1
L = gg (e2 x+ x + ) + e2 x+ x .
2
(2.13)
i
1 h
L = P + x + P x + + + 00 2P + P + e2 (2P x+ 2P + x + 2 + 2 )
2h
h01
.
+ 00 (P + x + P x+ + )
(2.14)
h
Compared to the AdSd model (2.7) in the present d = 3 case there is no transverse
degrees of freedom, and because of the WZ coupling the Lagrangian (2.14) does not
contain x+ x term. Fixing the light-cone gauge as in (2.8) and integrating over P
(getting again (2.9)) we find (l = 1)
h01
e2
.
L = + (2 + 2 ) + e2 x + (p+ x + )
2p
p
(2.15)
Using the expression for x which is implied by the constraint we can rewrite this
Lagrangian as
h01 +
,
L = + P + (p x + )
p
(2.16)
(2.17)
This is to be compared with the pure metric case (2.12) which for the AdS3 gives
P =
e2 2 2
( + ) .
2p+
(2.18)
1
2
p+
d P2 + e2/ x2 + e/ (2 + 2 ) ,
h
where 4 2 = R2 T0 = R2 /2 .
8
(3.1)
3.1
A quark source is represented as a static open string stretched from the horizon =
to the boundary = + of AdS space. For static solutions, P = = 0 and the
classical equations reduce to those extremizing the Hamiltonian:
(e2/ x ) = 0 ,
(3.2)
1 2/ 2
1
= 0.
e
x + e/ 2 (e/ )
(3.3)
(3.4)
where T is an integration constant. Note that for static solutions, the constraint (2.10),
implies that there is no extension in x . For an isolated quark,
i.e.
x = x P + ,
we also want no extension in x , so, in this case, we have T = 0.
If we define = e/ , the second equation (3.3) implies
T2
2
+ 2 2 = C ,
2 2
(3.5)
where C is a constant of integration. The left hand side is 1/ 2 times the density of P ,
so we may identify C = p / 2 p+ . For T = 0, one can trivially integrate the equation
to determine
2p
=
| +|,
2
p
(3.6)
where we have fixed the integration constant so that the end at = 0 is on the horizon
( = 0). The other end is at = p+ , so we find
mq
2|p+ p | = 2 max .
(3.7)
If that end reaches the boundary, max , implying an infinite mass for the quark
source, to be expected for a point charge.
3.2
A quark-antiquark source corresponds to an open string with both ends on the boundary
separated by the distance L. In this case T 6= 0 and there is a minimum value min > 0
of for an interior point of the string. For reasons of symmetry we shift the range to
mark this point by = 0: p+ /2 < < p+ /2. min occurs at the point where = 0, so
C = T2 /2 2 2min . Thus we have
mqq
2|p+ p | =
9
T p+
.
min
(3.8)
= T
p+
2
d
= 2
2
+
p2
2
=
min
min
2 /2min 1/
d
max /min
max
4 (3/4)(1/2)
, (3.9)
min
(1/4)
1/
where the last form is the limit as max . Similarly, we obtain as a function of
by direct integration leading to
p+
2
d
(min )3/2 Z max /min
q
,
T
1
1/
2|p+ p | = 2 min
max /min
(3.10)
1/
(3/4)(1/2)
4 max 4 min
,
(1/4)
(3.11)
where the last form gives the non-vanishing terms of the behavior as max . We
finally eliminate min in favor of L to reach the final result:
16 2 (3/4)2(1/2)2
2|p+ p | 4 max
L
(1/4)2
4 2 (2)3
= 4 max
.
L (1/4)4
(3.12)
1 2
2
P
+
e
e
+
e
A
,
2p+
10
(4.1)
1
,
X li j2 + (p+ 2 2)2 + 4p+ i li j j ,
(4.2)
4
where 2 i i and li j is the angular momentum operator of the su(4) algebra (for details
and are the bosonic conjugate momentum operators as in
see [36]). The P = (P, P)
(2.7). The odd part of the phase space is represented by i , i considered as fermionic
coordinates and i , i considered as fermionic momenta. Note that the Hamiltonian
does not depend on the fermionic variables i and i (present in the light-cone gauge
formulation of 4-dimensional N = 4 SYM theory [43]) but they will appear in the phase
space Lagrangian as i i + i i .
The canonical operator commutation relations are10
AX
[P, x] = i ,
{i , j } =
x] = i ,
[P,
1 i
,
p+ j
[, ] = i ,
{ i, j } =
1 i
.
p+ j
(4.3)
(4.4)
The operator A is equal to zero only for massless representations which are irreducible
representations of the conformal algebra [42, 46] (so(5, 2) in the case of AdS5 space). The
important property of the operator X (4.2) is that its eigenvalues are equal to squares
of integers for all states of type IIB supergravity compactified on S 5 . This fact plays
an important role in formulating the AdS/CFT correspondence for chiral primary states
[36]. The relation (4.2) then implies that the operator A is never equal to zero and
thus the scalar fields [47] as well as all other modes [36] of S 5 compactified type IIB
supergravity have equations of motion which are not conformally invariant. We expect
that this property of the operator X should have a string-theory analog.
The light-cone gauge phase space Lagrangian for the superparticle in AdS5 S 5 is
obtained from the Hamiltonian (4.1) in the usual way
i
L = P x + + PM u M + p+ (i i + i i + i i + i i ) + P .
(4.5)
2
uM is a unit 6-d vector used to parametrize S 5 . Note that here (and in the string case) we
treat x , P + = p+ separately from the rest of the phase space variables; p+ is conserved,
while x satisfies the equation
1
x = + P .
(4.6)
p
5
5.1
Let us first recall the form of the superstring action in the -symmetry light-cone gauge
[11]. It is formulated in terms of 10 bosonic coordinates (x , x, x
; , y A) (y A are 5 independent coordinates of S 5 ) and 16 fermionic coordinates (i , i , i , i ) which transform
10
Note that in the rest of the paper the brackets will stand for the classical Poisson brackets, and
thus there will be no is in similar relations.
11
gg e x+ i i + i i + i i + i i + ie2 x+ ( 2 )2 ,
2
LW Z = e2 x+ i CijU ( j i 2e j x) + h.c. ,
(5.2)
where
D y A = y A 2ii (V A )i j j e2 x+ .
(5.3)
(5.4)
Here GAB and (V A )i j are the metric tensor and the Killing vectors of S 5 respectively, i.e.
this Lagrangian corresponds to the following parametrization of the metric of AdS5 S 5
This form of the superstring action (which we shall call intermediate) is most convenient for deriving other forms which differ by the way one chooses the bosonic coordinates
that parametrize AdS5 S 5 (see [11] and Appendices B and C).
Another useful form is found by using a 6-d unit vector uM to parametrize S 5 , i.e.
by replacing (5.4) by
ds2 = e2 dxa dxa + d2 + duM duM ,
uM uM = 1 .
(5.5)
Lkin = gg e2 ( x+ x + x x) + + D uM D uM
2
2
i
h
i 2
gg e x+ i i + i i + i i + i i + ie2 x+ ( 2 )2 ,
2
j
M
j
LW Z = e2 x+ i M
ij u ( i 2e x) + h.c. ,
(5.6)
(5.7)
where
1
(5.8)
RM = M N uN .
2
The parametrization using uM is the most convenient one for the discussion of the superparticle in AdS5 S 5 [36, 48], and is well-suited for the harmonic decomposition of the
light-cone superfield of type IIB supergravity into the Kaluza-Klein modes [36]. Other
forms of the superstring action using 4+6 Cartesian coordinates for AdS5 S 5 (see
(1.3) and Appendix C) are directly related to (5.6),(5.7) by a coordinate transformation
(Y M = e uM , |Y | = e ).
D uM = uM 2ii (RM )i j j e2 x+ ,
12
The superstring Lagrangian (5.1),(5.2) and all of its other forms mentioned above
can be represented in the following way
L = L1 + L2 + L3 ,
(5.9)
1
1
L1 = h x+ x + x+ A + h x+ x+ B h gAB D y A D y B ,
2
2
(5.10)
1
L2 = h x x + x C ,
2
(5.11)
1
L3 = h e2 + T .
2
(5.12)
Here x = (x, x) ,
gAB e2 GAB ,
and h is defined as in (2.5), i.e.
h gg e2 ,
D y A y A + F A x+ ,
(5.13)
(5.14)
i
A = h (i i + i i ) + 1 e2 i CijU (j i 2e j x) + h.c. ,
2
B = e2 ( 2 )2 ,
C x = i 21 e3 x+ i CijU j ,
(5.16)
C x = (C x ) ,
F A = 2ie2 i (V A )i j j ,
(5.17)
(5.18)
T = e2 x+ i CijU j + h.c. .
5.2
(5.15)
(5.19)
L
,
x a
L
,
13
PA =
L
,
y A
(5.20)
we get
= h00 e2 + h01 e2 + ,
(5.21)
P + = h00 x + h01 x+ ,
(5.22)
P = h00 x h01 x + C 0 ,
(5.23)
P A = h00 y A h01 yA + F A P + ,
(5.24)
P = h00 x h01 x + A0 BP + + PA F A .
(5.25)
1 h +
2P P + 2e4 x+ x
2h00
h01 +
1
(P x + P x+ + PA yA ) 00 (P + + h01 x+ )A0 + x+ A1 ,
(5.26)
00
h
h
1
h01
= P x + 00 ((P C 0 )2 + e4 x2 ) + 00 (P C 0 ) x + x C1 , (5.27)
2h
h
01
1
h
= + 00 e2 (2 + 2 ) + 00 + T .
(5.28)
2h
h
+
L2
L3
P + = p+ .
(5.29)
Using these gauge conditions in the action and integrating over P we get the expression
for h00
h00 = p+ .
(5.30)
Inserting this into (5.26), (5.27) (5.28) we get the following general form of the phase
space light-cone Lagrangian
L1 = PA y A
1 AB
4
A B
+2
+ A
g
P
g
+
p
B
2p
F
P
A PB + e
AB y
A
2p+
h01 +
(p x + PA yA ) + A0 ,
p+
L2 = P x
L3 =
1
h01
2
4 2
(P
+
e
x
P x ,
2p+
p+
1 2 2 2
h01
e
(
+
.
2p+
p+
14
(5.31)
(5.32)
(5.33)
In deriving these expressions we used the fact the functions C , T given in (5.17),(5.19)
are equal to zero in the light-cone gauge (5.29). This general form of the phase space
Lagrangian can be used to derive explicit forms of Lagrangians corresponding to different choices of bosonic coordinates: one needs only to insert the appropriate functions
A0 , B, and F A . For the intermediate case (5.1),(5.2) these functions are given by
(5.15),(5.16),(5.18) so that we get
i
L = P x + + PA y A + p+ (i i + ii + i i + i i )
2
h
i
1
2
4 2
2
2
2 + li 2 + GAB yA yB + p+2 ( 2 )2 + 4p+ i li j j )
P
+
e
x
+
e
(
+
j
2p+
)
+ e2 i CijU (j i 2e j x) + e2 i CUij (j + i 2e j x
i
h01 h +
+ PA yA + i p+ (i i + i i + i i + i i ) .
p
x
+
P
x
p+
2
(5.34)
(5.35)
li j2 li j lj i .
(5.36)
By applying the coordinate transformation the above Lagrangian can be rewritten in the
form corresponding to the case (5.6),(5.7) in which the S 5 part is parametrized by the
unit 6-d vector uM
i
L = P x + + PM u M + p+ (i i + i i + i i + i i )
2
h
i
1
2
4 2
2
2
2 + li 2 + uM uM + p+2 ( 2 )2 + 4p+ i li j j )
P
+
e
x
+
e
(
+
j
2p+
)
+ e2 i yij (j i 2e j x) + e2 i y ij (j + i 2e j x
i
h01 h +
+ PM uM + i p+ (i i + i i + i i + i i ) ,
p
x
+
P
x
p+
2
(5.37)
M
yij M
= CijU ,
ij u
(5.38)
i
li j = (M N )i j uM P N .
(5.39)
2
Taking into account the constraint uM P M = 0 (see (5.59)) we get li k lk j = 14 P M P M i j .
The above Lagrangian gives the Hamiltonian
15
H = P ,
P =
d P ,
(5.40)
i
1 h 2
4 2
2
2
2 + li 2 + uM uM + p+2 ( 2 )2 + 4p+ i li j j )
P
+
e
x
+
e
(
+
j
2p+
) .
+ e2 iyij (j i 2e j x) + e2 i y ij (j + i 2e j x
(5.41)
P =
(5.42)
As usual, this constraint allows one to express the non-zero modes of the bosonic coordinate x in terms of the transverse physical ones.
It is easy to see that in the particle theory limit the superstring Hamiltonian (5.41)
reduces to the superparticle one in (4.1). The latter was found in [36] by applying the
direct method of constructing relativistic dynamics [45] based on the symmetry algebra.11
The present discussion thus provides a self-contained derivation of the light-cone gauge
superparticle action from the covariant one.
5.3
Equations of motion
The equations of motion corresponding to the phase space superstring Lagrangian (5.37)
take the following form
x =
1
P,
p+
x =
1
P ,
p+
e2
= + ,
p
1
4
(e
x
i
2 (e3 i y ij j ) ,
p+
1
) + i 2 (e3 i yij j ) ,
P = + (e4 x
p
P =
(5.43)
(5.44)
(5.45)
2 4 2
e2 2 2
1
2
i 2
M M
+2 2 2
+
i j
(e
)
e
x
+
l
+
u
+
p
(
)
+
4p
i
j
j
p+
p+
p+
) ,
+ e2 iyij (2j 3i 2e j x) + e2 i y ij (2j + 3i 2e j x
(5.46)
u M =
e2 M
P ie2 i (M N )i j j uN ,
+
p
(5.47)
11
Strictly speaking, the string Hamiltonian (5.41) reduces to (4.1) modulo terms proportional to 2
and some constant. This difference is related to the fact that here we are considering the classical string
Hamiltonian, i.e. ignore operator ordering, while the particle Hamiltonian (4.1) is defined in terms of
quantum operators.
16
1
e2
P M = + uM P N P N + + v M N (e2 uN ) ie2 i (M N )i j j P N
p
p
j
j
)
+ e2 v M N i N
) + e2 v M N i (N )ij (j + i 2e j x
ij ( i 2e x
(5.48)
(5.49)
i
i = + (e2 y ij j ) ,
p
(5.50)
(5.51)
i
i = + (e2 yij j ) ,
p
i
h
i
2i
) ,
= e2 i 2 i + (l)i + + y ij (j + i2 2e j x
p
p
h
i
2i
i
= e2 i 2 i + + (l)i + + yij (j i2 2e j x) .
p
p
(5.52)
Here we defined
v M N M N uM uN ,
(5.53)
X = [X , P ] ,
(5.54)
and, as previously, do not distinguish between the upper and lower indices M, N, i.e. use
the convention PM = P M . These equations can be written in the Hamiltonian form. Introducing the notation X for the phase space variables (P , x , , , P M , uM , i , i , i , i )
one has the Hamiltonian equations
where the phase space variables satisfy the (classical) Poisson-Dirac brackets
[ P(), x( ) ] = (, ) ,
[ P(),
x( ) ] = (, ) ,
[ (), ( ) ] = (, ) , (5.55)
[P M (), uN ( )] = v M N (, ) ,
{i (), ( )} = + i (, ) ,
{i (), ( )} = + i (, ) ,
(5.57)
p
p
1
1 i
1
1 i
i
i
, [x
i , [x
, [x
i ,
(5.58)
[x
0 , i ] =
0 , ] =
0 , i ] =
0 , ] =
+
+
+
2p
2p
2p
2p+
where x
0 is the zero mode of x . All the remaining brackets are equal to zero (with ex
+
ception of [p , x0 ] = 1). The structure of (5.56) reflects the fact that in the Hamiltonian
formulation the constraint uM uM = 1 should be supplemented by the constraint
uM P M = 0 .
(5.59)
These are second class constraints and the Dirac procedure leads then to the classical
Poisson-Dirac brackets (5.56). To derive (5.57),(5.58) one is to take into account that
the Lagrangian (5.37) has the following second class constraints
i
i
pi + p+ i = 0 ,
p i + p + i = 0 ,
(5.60)
2
2
where pi , pi are the canonical momenta of fermionic coordinates. The same constraints
are found for the fermionic coordinates i , i . Starting with the Poisson brackets
17
{pi , j }P.B. = ji ,
[p+ , x
{pi , j }P.B. = ij ,
0 ]P.B. = 1 ,
one gets then the Poisson-Dirac brackets given in (5.57),(5.58).
5.4
(5.61)
Finally, let us note that our results for the AdS5 S 5 string can be generalized in a
rather straightforward way to the case of superstring in AdS3 S 3 space with RR 3form background (for the corresponding covariant GS action see [49]). Our light-cone
gauge action is written in the form which allows a straightforward generalization to the
AdS3 S 3 case: one is just to do a dimensional reduction.
To get the Lagrangian and Hamiltonian P in AdS3 S 3 case we are to set x =
0, P = 0 in (5.37), (5.41). Now instead of su(4) so(6) we have so(4) which is
decomposed into su(2) and su(2).
p+
2
The (minus) Hamiltonian density is given by (cf. (2.18))
P =
where
(5.62)
(5.63)
i
e2 h 2 2
i 2
M M
+2 2 2
+
i j
+ e2 ( i yij j + h.c.) , (5.64)
+
2l
+
u
+
p
(
)
+
4p
i
j
j
2p+
yij = Cik
( M )k j uM ,
M = ( 1 , 2 , 3 , iI) .
(5.65)
and M, N = 1, 2, 3, 4. The expression for orbital part of SU(2) generators li j takes the
form
i
1
li j = ( M N )i j uM P N ,
( M N )i j = ( M )i k (
N )k j (M N) ,
(5.66)
2
2
where
M = ( 1 , 2 , 3 , iI). Note that in this case one has the relation 2li2j = P M P M ,
and this explains the factor of 2 in front of li2j in (5.64). The variation over h01 gives the
constraint
i
(5.67)
p+ x + + PM uM + p+ (i i + i i + i i + i i ) = 0 .
2
An interesting feature of the Hamiltonian (5.64) is that e2 factors out (cf. (5.41)). Note
also that the dependence on matrix Cij can be eliminated by the redefinition of the
fermionic coordinates i Cij j .
18
The Noether charges play an important role in the analysis of the symmetries of dynamical systems. The choice of the light-cone gauge spoils manifest global symmetries,
and in order to demonstrate that these global invariances are still present one needs to
find the Noether charges which generate them.12 These charges play a crucial role in
formulating superstring field theory in the light-cone gauge (see [50, 51]).13
The Noether charges for a superparticle in AdS5 S 5 were found in [36]. These
charges are helpful in establishing a correspondence between the bulk fields of type IIB
supergravity and the chiral primary operators of the boundary theory in a manifestly
supersymmetric way. Superstring Noether charges should thus be useful for the study
of the AdS/CFT correspondence at the full string-theory level.
In the light-cone formalism the generators (charges) of the basic psu(2, 2|4) superalgebra can be split into two groups:
Q+i , Q+ , S +i , S + , D, J + , J xx ,
P + , P, P , J +x , J +x , K + , K, K,
i
i
(6.1)
P , J x , J x , K , Qi , Q
i , S , Si ,
(6.2)
which we shall refer to as dynamical generators (see also [11]). The kinematic generators
have positive or zero J + charges, while dynamical generators have negative J + charges.
For x+ = 0 the kinematic generators in the superfield realization are quadratic in the
physical string fields (i.e. they have the structure J = J1 + x+ J2 + x+2 J3 where J1 is
quadratic but J2 , J3 contain higher order terms in second-quantized fields), while the
dynamical generators receive higher-order interaction-dependent corrections. The first
step in the construction of superstring field theory is to find a free (quadratic) superfield
representation of the generators of the psu(2, 2|4) superalgebra. The charges we obtain
below can be used to obtain (after quantization) these free superstring field charges.
6.1
As usual, symmetry generating charges can be obtained from conserved currents. Since
currents themselves may be helpful in some applications, we shall first derive them
starting with the -symmetry gauge fixed Lagrangian in the form given in (5.6),(5.7).
12
In what follows currents and charges will mean both bosonic and fermionic ones, i.e. will
include supercurrents and supercharges.
13
Let us note in passing that the development of the light-cone string field theory approach in the
case of the AdS5 S 5 background may be useful in the context of AdS/CFT correspondence. One
striking feature of the light-cone closed superstring field theory actions in the flat space case is that
their interaction part contains only cubic vertices. Same may happen to be true also in the case of the
AdS5 S 5 background, and that may have important implications for establishing correspondence with
the SYM theory in the light-cone gauge framework.
19
To obtain the currents we shall use the standard Noether method (see, e.g., [52]) based
on the localization of the parameters of the associated global transformations. Let be
a parameter of some global transformation which leaves the action invariant. Replacing
it by a function of worldsheet coordinates , , the variation of the action takes the form
S =
d2 G ,
(6.3)
where G is the corresponding current. Making use of this formula, we shall find below
those currents which are related to symmetries that do not involve compensating symmetry transformation. The remaining currents will be found in the next subsection
starting from the action (5.37) where both the -symmetry and the bosonic light-cone
gauges are fixed.
Let us start with the translation invariance xa = a . Applying (6.3) to the Lagrangian (5.6),(5.7) gives the translation currents
P + = gg e2 x+ ,
P = gg e2 x i 2e3 i y ij j x+ ,
P = gg e2 x + i 2e3 i yij j x+ ,
P = gg e2 x + F M D uM
(6.4)
(6.5)
(6.6)
i 2 i
gg e i + i i + i i + i i + 2ie2 x+ ( 2 )2
2
+ e2 i yij ( j i 2e j x) + e2 i y ij ( j + i 2e j x) ,
(6.7)
where
F M ii (M N )i j j e2 uN .
(6.8)
x = J x+ ,
x = J x ,
x = Jx+ ,
x = J x ,
(6.9)
J +x = gg e2 (x+ x x x+ ) ix+ e3 x+ i y ij j ,
(6.10)
(6.11)
J +x = x+ P xP + ,
(6.12)
Some of the remaining bosonic currents can be expressed in terms of supercurrents. The
invariance with respect to the super-transformations
i = i ,
i
i
x = i i i i ,
2
2
i = i ,
(6.13)
Q+i = gg e2 i x+ i e2 y ij j x+ ,
Q+
= gg e2 i x+ i e2 yij j x+ .
i
(6.14)
(6.15)
The invariance of the action (5.6),(5.7) with respect to the rotation of (super)coordinates
in the (x+ , x ) plane
x = e x ,
(i , i , i , i ) = e/2 (i , i , i , i ) ,
(6.16)
J + = gg e2 (x+ x x x+ )
i 2 + i
gg e x i + i i + i i + i i + 2ie2 x+ ( 2 )2
2
+ x+ e2 i yij ( j i 2e j x) + x+ e2 i y ij ( j + i 2e j x)
1
1 2
e x+ i yij j e2 x+ i y ij j .
2
2
(6.17)
x = ei x ,
i = e 2 i ,
i = e 2 i ,
i = e 2 i ,
i = e 2 i ,
i
J xx = gg e2 (x x x x) +
gg x+ (i i i i )
2
1
1 2
e x+ i yij (j i2 2 j x) + e2 x+ i y ij (j + i2 2j x) (6.19)
,
2
2
which can be rewritten in terms of translation currents as follows
i
i
i
J xx = xP xP i Q+
+ i Q+i + (i i + i i )P +
i
2
2
2
The invariance with respect to the dilatations
xa = e xa ,
= ,
(i , i , i , i ) = e/2 (i , i , i , i ) ,
21
(6.20)
(6.21)
D = gg e2 xa xa + F M D uM
i 2 + i
gg e x i + i i + i i + i i + 2ie2 x+ ( 2 )2
2
+ x+ e2 i yij ( j i 2e j x) + x+ e2 i y ij ( j + i 2e j x)
2
2
e x+ i yij (j i2 2e j x)
e x+ i y ij (j + i2 2e j x) .
2
2
(6.22)
i
i
gg i Q+
i Q+i .
i
2
2
i
The invariance with respect to the SU(4) rotations ( i = 0)
D = xa P a +
y ij = i l y lj + j l y il ,
i = i j j ,
1
uM = i j (M N )j i uN ,
2
i.e.
i = j j i ,
i = i j j ,
i = j j i ,
(6.23)
(6.24)
(6.25)
J i j = i gg (M N )i j uM D uN + [i j + i j ji (l l + l l )]P +
2
4
1
1
i e2 x+ (i yjl l ji k ykl l ) + i e2 x+ (j y il l ji k y kl l ) .(6.26)
4
4
6.2
In the previous section we have found (super)currents starting with the -symmetry lightcone gauge fixed action given in (5.6),(5.7). These currents can be used to find currents
for the action where both the fermionic -symmetry and the bosonic reparametrization
symmetry are fixed by the light-cone type gauges (5.37). To find the components of
currents in the world-sheet time direction G 0 one needs to use the relations (5.21)(5.25)
for the canonical momenta and to insert the light-cone gauge conditions (5.29) and (5.30)
into the expressions for the currents given in the previous subsection. The charges are
then given by
G=
d G 0 .
(6.27)
Let us start with the kinematic generators (charges) (6.1). The results for the currents
imply the following representations for some of them
P =
P,
P =
22
P ,
P + = p+ ,
(6.28)
J +x =
J +x =
x+ P xp+ ,
+i
+ i
Q+
i
p ,
x+ P xp+ ,
(6.29)
p+ i .
(6.30)
Note that these charges depend only on the zero modes of string coordinates. In (6.28)
(6.30) the integrands are G 0 parts of the corresponding currents in world-sheet time
direction: P 0 , P 0 , J +x0 , J +x0 , Q+i0 , Q+0
and P +0 = p+ . The components of currents
i
1
in the world-sheet spatial direction G can be found simply by using the conservation
laws, the expressions for G 0 and the equations of motion (5.43)(5.52). In this way we
obtain
P1 =
3 ij
1 4
2e i y j ,
e
x
+
i
p+
1
i 2e3 i yij j ,
P 1 = + e4 x
p
(6.31)
i 2
e yij j .
p+
(6.32)
The remaining kinematic charges depend on non-zero string modes and are given by
J +x1 = x+ P 1 ,
J +x1 = x+ P 1 ,
x
x
J ij
J
Q+i1 =
i 2 ij
e y j ,
p+
Q+1
i =
i
i
xP xP p+ 2 + p+ 2 ,
2
2
Z
1
= li j + p+ i j + p+ i j ji p+ (2 + 2 ) ,
4
=
D=
x+ P + x p+ + xP + xP .
x+ P x p+ ,
(6.33)
(6.34)
(6.35)
(6.36)
The derivation of the remaining charges can be found in Appendix D. The expressions
for the conformal supercharges are given by
Si+ = Si+ |x+ =0 ix+ Q
i ,
S +i = S +i |x+ =0 + ix+ Qi ,
(6.37)
1
e p+ i + ip+ i x ,
2
S +i |x+ =0 =
1
e p+ i ip+ i x .
2
(6.38)
Q
i
Qi
e
Pi + ii p+ 2 i + 2(l)i + yij (j i 2e j x) ,
(6.39)
2
Z
e
i+
) , (6.40)
i i p+ 2 i + 2(l)i y ij (j + i 2e j x
= P
2
23
where
(l)i j lj i ,
(l)i li j j .
(6.41)
(6.42)
(6.43)
1
x)p+ ,
(e2 + 2x
2
Z
1
= e2 P + x(x p+ + xP +
2
Z
1
= e2 P + x(x p+ + xP
2
K |x+ =0 =
K|x+ =0
x+ =0
K|
(6.44)
i + 2 i + 2
p + p ) i Si+0 |x+ =0 , (6.45)
2
2
i + 2 i + 2
p p ) + i S +i0 |x+ =0 . (6.46)
2
2
Here the densities of conformal supercharges Si+0 |x+ =0 , S +i0 |x+ =0 are given by the integrands in (6.38).
Note that the G|x+ =0 parts of the kinematic charges can be obtained from the superparticle ones simply by replacing the particle coordinates by the string ones. The
remaining dynamical generators J x , J x , S i , Si , K can be found by using the expression found above and applying the commutation relations of psu(2, 2|4) superalgebra.
The structure of the psu(2, 2|4) generators we have presented in this section is, of
course, more complicated than found in the flat space case. But still there are some interesting simplifications which have an algebraic origin. A remarkable feature of psu(2, 2|4)
as compared to the Poincare superalgebra is that in order to find all of its dynamical
generators it is sufficient to know only the kinematic ones (6.1) and the Hamiltonian P
(5.40): the dynamical generators are obtained using commutation relations between the
kinematic generators and the Hamiltonian P .
Acknowledgments
The work of R.R.M. and A.A.T. was supported by the DOE grant DE-FG02-91ER40690 and the INTAS project 991590. R.R.M. is also supported by the RFBR Grant
No.99-02-16207. A.A.T. would like to acknowledge also the support of the EC TMR
grant ERBFMRX-CT96-0045 and the PPARC SPG grant PPA/G/S/1998/00613, and
acknowledges also the hospitality of the Aspen Center for Physics and the CERN Theory
Division during the completion of this paper. The work of C.B.T. was supported by the
Department of Energy under Grant No. DE-FG02-97ER-41029.
24
Appendix A
Notation
In the main part of the paper we use the following conventions for the indices:
a, b, c = 0, . . . , 3
A, B, C = 1, . . . , 5
A , B , C = 1, . . . , 5
M, N, K, L = 1, . . . , 6
i, j, k, l = 1, . . . , 4
, = 0, 1
1
x, x = (x1 ix2 ) .
2
(A.1)
(A.2)
A Ax = Ax ,
(A.3)
We use the notation x to represent (x, x) with the following summation rule
x z = x
z + xz .
(A.4)
x x
(A.5)
M
M
ij = ji ,
(M )ij (M
ij ) .
(A.7)
The SO(5) Dirac and charge conjugation matrices can be expressed in terms of the M
matrices as follows
Cij = 6ij .
(A.8)
1
M kl
M
ij = ijkl ( ) ,
2
The matrices M N are defined by
M kl
l k
k l
M
ij ( ) = 2(i j i j ) .
(A.9)
1
(M N )i j (M )il N
lj (M N) ,
2
(A.10)
(M N )i j (M N )k l = 2ji lk 8li jk .
(A.11)
so that
We use the following hermitian conjugation rule for the fermionic coordinates
i = i ,
i = i ,
(A.12)
Appendix B
2 i i .
(A.13)
GAB = eA
eA ,
A B
GAB =
eA
=
A
sin2 |y|
(AB nA nB ) + nA nB ,
|y|2
nA nA = 1 ,
sin |y| A
(A nA nA ) + nA nA ,
|y|
nA
yA
,
|y|
|y| =
(B.1)
(B.2)
y A y A .
(B.3)
We use the convention y A = AA y A and the same for nA . The S 5 Killing vectors V A
and V A B corresponding to the five translations and ten SO(5) rotations respectively
are given by
V A B = y A yB y B yA .
(B.4)
(B.5)
(V A )i j yA = ( A B )i j V A B + ( A )i j V A .
4
2
The commutation relations of so(6) algebra then include
[V A , V B ] = V A B ,
(B.6)
[V A , V B C ] = A B V C A C V B .
26
(B.7)
(B.8)
This matrix is related to the standard charge conjugation matrix Cij (which has the
uA = sin |y|nA ,
u6 = cos |y| ,
(B.9)
(B.10)
(B.12)
y ij (M )ij uM ,
y il ylj = ji .
(B.13)
Appendix C
yij = y ij ,
yij = yji .
(B.14)
The light-cone gauge superstring action of [11] can be written in several equivalent forms
corresponding to different parametrizations of AdS5 S 5 space. The two forms corresponding to 5+5 parametrizations (5.1),(5.2) and (5.6),(5.7) were discussed in Section
5. Here we shall review the Lagrangians [11] for the two standard 4+6 coordinate
choices and present their phase space counterparts.
27
Choosing the 10 Cartesian coordinates (xa , Y M ) such that the metric of AdS5 S 5
takes the form
ds2 = Y 2 dxa dxa + Y 2 dY M dY M ,
Y 2 = |Y |2 = Y M Y M ,
(C.1)
one can transform the bosonic coordinates in the intermediate form of the superstring
Lagrangian (5.1),(5.2) to obtain its 4+6 form [11]
L = Lkin + LW Z ,
i
h
1
Lkin = gg Y 2 ( x+ x + x x) + Y 2 D Y M D Y M
2
i
h
i 2
gg Y x+ i i + i i + i i + i i + iY 2 x+ ( 2 )2 ,
2
M
LW Z = |Y | x+ i M
( j i 2|Y | j x) + h.c. .
ij Y
(C.2)
(C.3)
Here
1
RM = M N Y N ,
(C.4)
2
and the matrices M , M N were defined in (A.6),(A.10).
It is easy to see that this Lagrangian can be represented in the same form as (5.9).
Applying the final result for the phase space Lagrangian given in (5.31)(5.33) we get
DY M = dY M 2ii (RM )i j j Y 2 dx+ ,
i
L = P x + PM Y M + p+ (i i + i i + i i + i i )
2
h
i
1
2
4
4 2
M M
2 +2 2 2
+
M
P
+
|Y
|
P
P
+
|Y
|
x
+
Y
Y
+
Y
(p
(
)
+
4ip
R
P
)
M M
M
2p+
i
h
M j
j
Y
(
i
2|Y
|
x
)
+
h.c.
+
|Y | iM
ij
i
h01 h +
M + i p+ (i i + i i + i i + i i ) .
p
x
+
P
x
+
P
Y
M
p+
2
(C.5)
The closely related choice of 4+6 coordinates is the one which makes explicit the hidden
spacetime conformal symmetry of AdS5 S 5 geometry, i.e. the one in which the metric
takes the conformally flat form
YM
.
(C.6)
Y2
The superstring Lagrangian expressed in terms of these conformally flat coordinates
is readily obtained from (C.2),(C.3) [11]
ds2 = Z 2 (dxa dxa + dZ M dZ M ) ,
ZM =
i
h
1
Lkin = gg Z 2 x+ x + x x + D Z M D Z M
2
i
h
i 2
gg Z x+ i i + i i + i i + i i + iZ 2 x+ ( 2 )2 ,
2
1 j
M
j
LW Z = |Z|3 x+ iM
ij Z ( i 2|Z| x) + h.c. ,
28
(C.7)
(C.8)
where
1
(C.9)
RM = M N Z N .
2
This Lagrangian has again the same form as in (5.9)(5.12), and its phase space counterpart is thus found from (5.26)(5.28)
DZ M = dZ M 2ii (RM )i j j Z 2 dx+ ,
i
L = P x + PM Z M + p+ (i i + i i + i i + i i )
2
h
i
1
2
4 2
M Z M ) + Z 2 (p+2 ( 2 )2 + 4ip+ RM PM )
P
+
P
P
+
Z
(
x
+
Z
M
M
2p+
h
i
M j
1 j
Z
(
i
2|Z|
+
|Z|3 iM
)
+
h.c.
ij
i
h01 h +
M + i p+ (i i + i i + i i + i i ) .
p
x
+
P
x
+
P
Z
M
p+
2
(C.10)
An interesting feature of this Lagrangian is that the squares of the AdS5 and S 5 momenta
enter exactly as in the flat space case.
Appendix D
Conformal supercharges
In order to find the conformal supercharges supplementing the generators given in Section
6, one needs to know the contribution of the compensating -symmetry transformation
given in [4]. Finding these compensating transformations directly is rather complicated.
Here we suggest an indirect method based on exploiting the known equations of motion
given in (5.43)(5.52) and the commutation relations of the psu(2, 2|4) superalgebra. Let
us demonstrate this procedure for the case of the conformal supercharges S +i .
In general, the charges depend on the time variables x+ = explicitly and also
through the dynamical variables, i.e. G = G(, X ( )). Because our equations of motion
dG
(5.43)(5.52) have a Hamiltonian form (5.54) we can rewrite the conservation law dx
+ = 0
as follows (as in (5.54)(5.58) here [ , ] stand for the classical Poisson bracket)
G
+ [G, P ] = 0 .
x+
This determines the explicit dependence on x+ , namely,
Gx+ = Gx+ =0 + x+ [P , Gx+ =0 ] +
x+2
[P , [P , Gx+ =0 ]] ,
2
(D.1)
(D.2)
(D.3)
and took into account that the expansion in x+ terminates at (x+ )2 order because the
psu(2, 2|4) superalgebra does not have generators with absolute value of the J + charge
higher than 1 (so that triple and higher commutators of P with any generator vanish).
29
For example, eq. (D.2) implies that the charges J +x , J + , D depend on x+ as follows
J +x = J +x |x+ =0 + x+ P ,
J +x = J +x |x+ =0 + x+ P ,
=J
|x+ =0 + x P ,
D = D|x+ =0 + x P .
(D.4)
(D.5)
These expressions match the ones given in (6.29),(6.34),(6.36). We also learn from (D.2)
and the commutation relations of psu(2, 2|4) that the dynamical generators and J i j , J xx
do not explicitly depend on x+ (cf. (6.33),(6.34)).
Let us now turn to the conformal supercharges. From (D.2) and commutation relations of psu(2, 2|4) we find
Si+ = Si+ |x+ =0 ix+ Q
i ,
and thus
Q
i |x+ =0 = Qi ,
[Si+ |x+ =0 , P ] = iQ
i ,
S +i = S +i |x+ =0 + ix+ Qi ,
(D.6)
Qi |x+ =0 = Qi ,
(D.7)
[S +i |x+ =0 , P ] = iQi .
(D.8)
Si+
Si+
(D.9)
(D.10)
Q+
i
Q1
i
e
+ 2
j i 2e j x) ,
Q0
=
P
+
(D.12)
i
+
2(l)
+
y
(
i
i
i
i
ij
i
2
e
ie2
ie2
= e3 i ( + x + i 2k y kll ) + + yij j P + + (e i ) + 2 2e j (lj i )1 , (D.13)
p
p
2p
where
i 2 M N i M N
e ( ) j u u .
(D.14)
2p+
These relations determine the supercharges. Finally, one can check that these charges
satisfy the commutation relations of the psu(2, 2|4) superalgebra. The corresponding
commutation relations are obtainable from the ones given in Section 3 of [11] by making
the following substitutions there: J i j iJ i j , P a P a , K a K a , S S.
(li j )1
30
References
[1] A.M. Polyakov, String theory as a universal language, hep-th/0006132. The wall
of the cave, Int. J. Mod. Phys. A14, 645 (1999) hep-th/9809057; String theory
and quark confinement, Nucl. Phys. Proc. Suppl. 68, 1 (1998) hep-th/9711002.
[2] J. Maldacena, The large-N limit of superconformal field theories and supergravity,
Adv. Theor. Math. Phys. 2, 231 (1998), hep-th/9711200. S.S. Gubser, I.R. Klebanov
and A.M. Polyakov, Gauge theory correlators from non-critical string theory,
Phys. Lett. B428, 105 (1998), hep-th/9802109. E. Witten, Anti-de Sitter space
and holography, Adv. Theor. Math. Phys. 2, 253 (1998), hep-th/9802150.
[3] M.B. Green and J.H. Schwarz, Covariant description of superstrings, Phys. Lett.
B136, 367 (1984); Nucl. Phys. B243, 285 (1984).
[4] R.R. Metsaev and A.A. Tseytlin, Type IIB superstring action in AdS5 S 5 background, Nucl. Phys. B533, 109 (1998), hep-th/9805028.
[5] R. Kallosh, J. Rahmfeld and A. Rajaraman, Near horizon superspace, JHEP
9809, 002 (1998) [hep-th/9805217].
[6] R.R. Metsaev and A.A. Tseytlin, Supersymmetric D3 brane action in AdS5 S 5 ,
Phys. Lett. B436, 281 (1998) [hep-th/9806095].
[7] I. Pesando, A kappa gauge fixed type IIB superstring action on AdS5 S 5 , JHEP
9811, 002 (1998), [hep-th/9808020]; All roads lead to Rome: Supersolvable and
supercosets, Mod. Phys. Lett. A14, 343 (1999) [hep-th/9808146].
[8] R. Kallosh and J. Rahmfeld, The GS string action on AdS5 S 5 , Phys. Lett.
B443, 143 (1998), hep-th/9808038.
[9] R. Kallosh and A.A. Tseytlin, Simplifying superstring action on AdS5 S 5 , JHEP
10, 016 (1998), hep-th/9808088.
[10] A.R. Kavalov and A.G. Sedrakian, Quantum Geometry Of Covariant Superstring
With N=1 Global Supersymmetry, Phys. Lett. B182, 33 (1986). P.B. Wiegmann,
Extrinsic Geometry Of Superstrings, Nucl. Phys. B323, 330 (1989).
[11] R.R. Metsaev and A.A. Tseytlin, Superstring action in AdS5 S 5 : -symmetry
light cone gauge, hep-th/0007036.
[12] S. Forste, D. Ghoshal and S. Theisen, Stringy corrections to the Wilson loop in N
= 4 super Yang-Mills theory, JHEP 08, 013 (1999), hep-th/9903042
[13] N. Drukker, D.J. Gross and A.A. Tseytlin, Green-Schwarz string in AdS5 S 5 :
Semiclassical partition function, JHEP 0004, 021 (2000) [hep-th/0001204].
[14] J. Maldacena, Wilson loops in large N field theories, Phys. Rev. Lett. 80, 4859
(1998), hep-th/9803002; S.-J. Rey and J. Yee, Macroscopic Strings as Heavy
Quarks of Large N Gauge Theory and Anti-de Sitter Supergravity, hep-th/9803001.
[15] I. Pesando, On the fixing of the kappa gauge symmetry on AdS and flat background: The lightcone action for the type IIB string on AdS5 S 5 , hep-th/9912284.
[16] M.B. Green and J.H. Schwarz, Supersymmetric String Theories, Phys. Lett.
B109, 444 (1982).
31
[17] R.E. Rudd, Light cone gauge quantization of 2-D sigma models, Nucl. Phys.
B427, 81 (1994) [hep-th/9402106].
[18] A.M. Polyakov, private communication (Jan. 2000).
[19] P. Goddard, J. Goldstone, C. Rebbi and C.B. Thorn, Quantum Dynamics Of A
Massless Relativistic String, Nucl. Phys. B56, 109 (1973).
[20] U. Lindstrom, B. Sundborg and G. Theodoridis, The Zero tension limit of the
superstring, Phys. Lett. B253, 319 (1991). J. Isberg, U. Lindstrom and B. Sundborg, Space-time symmetries of quantized tensionless strings, Phys. Lett. B293,
321 (1992) [hep-th/9207005]. J. Isberg, U. Lindstrom, B. Sundborg and G. Theodoridis, Classical and quantized tensionless strings, Nucl. Phys. B411, 122 (1994)
[hep-th/9307108].
[21] N. Berkovits, Covariant quantization of the superstring, [hep-th/0008145].
[22] M.A. Vasiliev, Consistent Equation For Interacting Gauge Fields Of All Spins
In (3+1)-Dimensions, Phys. Lett. B243, 378 (1990); Properties of equations of
motion of interacting gauge fields of all spins in (3+1)-dimensions, Class. Quant.
Grav. 8, 1387 (1991).
[23] M.A. Vasiliev, Higher-spin gauge theories in four, three and two dimensions, Int.
J. Mod. Phys. D5, 763 (1996) [hep-th/9611024].
[24] P. Claus, M. Gunaydin, R. Kallosh, J. Rahmfeld and Y. Zunger, Supertwistors as
quarks of SU(2, 2|4), JHEP 9905, 019 (1999) [hep-th/9905112].
[25] I. Bandos, J. Lukierski, C. Preitschopf and D. Sorokin, OSp supergroup manifolds,
superparticles and supertwistors, Phys. Rev. D61, 065009 (2000) [hep-th/9907113].
[26] F. Lund and T. Regge, Unified Approach To Strings And Vortices With Soliton
Solutions, Phys. Rev. D14, 1524 (1976).
[27] B.M. Barbashov and V.V. Nesterenko, Relativistic String Model In A Space-Time
Of A Constant Curvature, Commun. Math. Phys. 78, 499 (1981).
[28] B.M. Barbashov and V.V. Nesterenko, Introduction To The Relativistic String
Theory, Singapore, World Scientific (1990) 249 p.
[29] H. J. De Vega and N. Sanchez, Exact integrability of strings in D-Dimensional
De Sitter space-time, Phys. Rev. D47, 3394 (1993). M. Zyskin, On gauge fields:
Strings duality as an integrable system, hep-th/0004105.
[30] E.S. Fradkin and R.R. Metsaev, Quantum R Matrix In The Relativistic String
Model In A Space Of Constant Curvature, Mod. Phys. Lett. A5, 1329 (1990).
[31] G. t Hooft, A Planar Diagram Theory For Strong Interactions, Nucl. Phys. B72,
461 (1974).
[32] C.B. Thorn, On The Derivation Of Dual Models From Field Theory, Phys. Lett.
B70, 85 (1977). On The Derivation Of Dual Models From Field Theory. 2, Phys.
Rev. D17, 1073 (1978).
[33] R. Brower, R. Giles and C.B. Thorn, The Spectrum Of QCD In The Limit Of An
Infinite Number Of Colors At Fixed Coupling, Phys. Rev. D18, 484 (1978).
32
[34] K. Bering, J.S. Rozowsky and C.B. Thorn, QCD fishnets revisited, Phys. Rev.
D61, 045007 (2000) [hep-th/9909141].
[35] H.B. Nielsen and P. Olesen, A Parton View On Dual Amplitudes, Phys. Lett. B32,
203 (1970). B. Sakita and M.A. Virasoro, Dynamical model of dual amplitudes,
Phys. Rev. Lett. 24 (1970) 1146.
[36] R.R. Metsaev, Light cone gauge formulation of IIB supergravity in AdS5 S 5
background and AdS/CFT correspondence, Phys. Lett. B468, 65 (1999) [hepth/9908114].
[37] L. Brink, P. Di Vecchia and P. Howe, A Locally Supersymmetric And Reparametrization Invariant Action For The Spinning String, Phys. Lett. B65, 471 (1976).
[38] A.M. Polyakov, Quantum geometry of bosonic strings, Phys. Lett. B103, 207
(1981).
[39] R. Tzani, Light Cone Gauge In Polyakovs Theory, Phys. Rev. D38, 3112 (1988).
C. Arvanitis and R. Tzani, Equivalence between light-cone and conformal gauge
in the two-dimensional string, hep-th/9602109. E. Smith, Light cone gauge for
(1+1) strings, Nucl. Phys. B382, 229 (1992) [hep-th/9201049].
[40] G.T. Horowitz and A.R. Steif, Strings In Strong Gravitational Fields, Phys. Rev.
D42, 1950 (1990).
[41] R.R. Metsaev, Light cone formalism in AdS spacetime, in Proc. XIV Int. Workshop on High Energy Physics and Quantum Field Theory (QFTHEP99, Moscow,
27 May - 2 June, 1999), Eds. B.B.Levchenko and V.I.Savrin, MSU-Press 2000, hepth/9911016.
[42] R.R. Metsaev, Light cone form of field dynamics in anti-de Sitter spacetime and
AdS/CFT correspondence, Nucl. Phys. B563, 295 (1999) [hep-th/9906217].
[43] L. Brink, O. Lindgren and B.E. Nilsson, N=4 Yang-Mills Theory On The Light
Cone, Nucl. Phys. B212, 401 (1983); The Ultraviolet Finiteness Of The N=4
Yang-Mills Theory, Phys. Lett. B123, 323 (1983). S. Mandelstam, Light Cone
Superspace And The Ultraviolet Finiteness Of The N=4 Model, Nucl. Phys. B213,
149 (1983). L. Brink, M.B. Green and J.H. Schwarz, Ten-Dimensional Supersymmetric Yang-Mills Theory With SO(8) Covariant Light Cone Superfields, Nucl.
Phys. B223, 125 (1983).
[44] A. Rajaraman and M. Rozali, On the quantization of the GS string on AdS5 S 5 ,
Phys. Lett. B468, 58 (1999) [hep-th/9902046].
[45] P. A. Dirac, Forms Of Relativistic Dynamics, Rev. Mod. Phys. 21, 392 (1949).
[46] R.R. Metsaev, All conformal invariant representations of d-dimensional anti-de
Sitter group, Mod. Phys. Lett. A10, 1719 (1995).
[47] H. J. Kim, L. J. Romans and P. van Nieuwenhuizen, The Mass Spectrum Of Chiral
N=2 D = 10 Supergravity On S 5 , Phys. Rev. D32, 389 (1985).
[48] R.R. Metsaev, IIB supergravity and various aspects of light-cone formalism in
AdS space-time, Talk given at International Workshop on Supersymmetries and
Quantum Symmetries (SQS 99), Moscow, Russia, 27-31 Jul 1999. hep-th/0002008.
33
[49] I. Pesando, The GS type IIB superstring action on AdS3 S 3 T 4 , JHEP 02, 007
(1999), hep-th/9809145. J. Rahmfeld and A. Rajaraman, The GS string action on
AdS3 S 3 with Ramond-Ramond charge, Phys. Rev. D60, 064014 (1999), hepth/9809164. J. Park and S. Rey, Green-Schwarz superstring on AdS3 S 3 , JHEP
01, 001 (1999), hep-th/9812062.
[50] M.B. Green, J. H. Schwarz and L. Brink, Superfield Theory Of Type II Superstrings, Nucl. Phys. B219, 437 (1983).
[51] M.B. Green and J. H. Schwarz, Superstring Field Theory, Nucl. Phys. B243, 475
(1984); Superstring Interactions, Nucl. Phys. B218, 43 (1983).
[52] S. Weinberg, The Quantum theory of fields. Vol. 1: Foundations, Cambridge, UK:
Univ. Pr. (1995) 609 p.
34