1101.5987v1 Saemann 2

a
r
X
i
v
:
1
1
0
1
.
5
9
8
7
v
1

[
h
e
p
-
t
h
]

3
1

J
a
n

2
0
1
1
Branes, Quantization and Fuzzy Spheres
Christian Smann and Richard J. Szabo

HeriotWatt University, Edinburgh, U.K.
E-mail: C.Saemann@hw.ac.uk , R.J.Szabo@hw.ac.uk
We propose generalized quantization axioms for NambuPoisson manifolds, which allow for a
geometric interpretation of n-Lie algebras and their enveloping algebras. We illustrate these ax-
ioms by describing extensions of BerezinToeplitz quantization to produce various examples of
quantum spaces of relevance to the dynamics of M-branes, such as fuzzy spheres in diverse di-
mensions. We briey describe preliminary steps towards making the notion of quantized 2-plectic
manifolds rigorous by extending the groupoid approach to quantization of symplectic manifolds.
Corfu Summer Institute on Elementary Particles and Physics Workshop on Noncommutative Field Theory
and Gravity
September 812, 2010
Corfu, Greece
Speaker.
Report numbers: HWM111 , EMPG1103

c Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence. http://pos.sissa.it/
Branes, Quantization and Fuzzy Spheres Christian Smann and Richard J. Szabo
Contents
1. Brane dynamics and fuzzy spheres 2
1.1 Generalized Nahm equations 2
1.2 n-Lie algebras 4
1.3 Geometrical meaning of n-Lie algebras 5
1.4 Quantum geometry of branes 5
2. Quantization 6
2.1 Axioms of quantization 6
2.2 Axioms of generalized quantization 7
3. BerezinToeplitz quantization 8
4. Quantization of spheres 9
4.1 Quantization of S
2
: The fuzzy sphere 9
4
10
3
10
4.4 Fuzzy scalar eld theory on S
4
10
5. Quantization of non-compact manifolds 11
5.1 Quantization of R
3
11
5.2 Quantization of hyperboloids 12
6. Gerbes and quantization of loop spaces 12
6.1 Quantization of 2-plectic manifolds 12
6.2 Transgression to loop space 14
1. Brane dynamics and fuzzy spheres
1.1 Generalized Nahm equations
A large variety of quantum eld theories arise in the low energy limits of string theory. Many
properties and phenomena of string theory are reected in these low energy descriptions. At the
same time, embedding eld theories in string theory provides interesting new points of view, lead-
ing to unexpected insights and conjectures. It also provides a means of realizing solitonic objects
of string theory, collectively called branes, in terms of conventional topological solitons of eld
theory, such as monopoles, vortices, and instantons. The dynamics of these objects describe the
nonperturbative regime of string theory.
2
For example, the low energy effective eld theory on a collection of coincident D3-branes in
Type IIB string theory is N = 4 supersymmetric YangMills theory in four dimensions, obtained
using T-duality as the dimensional reduction of ten-dimensional N = 1 supersymmetric Yang
Mills theory. A class of BPS solutions to the supersymmetric YangMills equations are magnetic
monopoles which obey the Bogomolnyi equations
F = D
with Higgs eld . These monopoles can be interpreted as D1-branes ending on the D3-branes.
Since these objects are BPS congurations, one can trivially form a stack of N D1-branes. From the
perspective of the D1-branes, the effective dynamics is described by the Nahm equations [1, 2, 3]
dX
i
ds
+
i jk
[X
j
, X
k
] = 0 ,
where X
i
(s), i = 1, 2, 3 are NN matrix elds describing the transverse uctuations inside the D3-
brane, and s denotes the zero-mode of the Higgs eld which can be identied with the worldsheet
spatial coordinate of the D1-branes [2]. These equations have a fuzzy funnel solution given by
X
i
(s) =
1
s

i
with
i
=
i jk
[
j
,
k
]
in terms of generators
i
of the Lie algebra su(2). Points in the worldsheet of the D1-brane are
thus blown up into fuzzy two-spheres S
2
with radius
1
s
. This describes the polarization of the
D1-branes which carry magnetic charge due to the anomalous coupling of the D3-branes to a
RamondRamond two-form gauge potential [4].
While there is an effective description of D-branes, a complete such theory is lacking for M-
branes. The situation just described has a natural lift to M-theory wherein M2-branes ending on
M5-branes are conjecturally described by Nahm-type equations. The analog of BPS equations on
a at M5-brane are given by
2
X
a
= 0 and
i
H
i jk
= 0 ,
where X
a
are coordinates along the membranes and H is the three-form eld strength on the world-
volume of the M5-brane. For a soliton solution we take
H
01i
=
i
and H
i jk
=
i jkl
l
,
which describes the M2-brane as a solitonic conguration inside the M5-brane. From the perspec-
tive of the M2-branes, one postulates the existence of four scalar elds X
i
, i =1, 2, 3, 4, representing
the transverse uctuations inside the M5-brane, and satisfying the BasuHarvey equations [5]
dX
i
ds
+
i jkl
[X
j
, X
k
, X
l
] = 0 ,
where the 3-bracket [, , ] describes a generalization of the notion of Lie algebra to an entity
called a 3-Lie algebra. Similarly to the D1D3 system, these equations have a solution
X
i
(s) =
1
2s
i
with
i
=
i jkl
[
j
,
k
,
l
] .
3
We interpret this solution again as a fuzzy funnel, but this time with fuzzy three-spheres S
3
of
radius
1
2s
.
One of the goals in the following will be to provide a geometrical meaning and proper deni-
tion of these noncommutative spheres. We will work from the fact that generalized Nahm equations
are built on n-Lie algebras. We now explain this concept in more detail.
1.2 n-Lie algebras
An n-Lie algebra is a complex vector space A equipped with a totally antisymmetric n-ary map
[, . . . , ] : A
n
A
satisfying the fundamental identity
_
a
1
, . . . , a
n1
, [b
1
, . . . , b
n
]
=
n
i=1
_
b
1
, . . . , b
i1
, [a
1
, . . . , a
n1
, b
i
], b
i+1
, . . . , b
n
for a
i
, b
j
A, which can be regarded as an n-Jacobi identity [6]. Gauge transformations are built
from inner derivations : A
(n1)
Der(A) dened for a
i
A via
a
1
a
n1 (b) := [a
1
, . . . , a
n1
, b] , b A .
The fundamental identity ensures that they form a Lie subalgebra g
A
of End(A). Note that xing
one slot of an n-Lie bracket reduces it to an (n1)-Lie bracket. The case n = 2 corresponds to the
usual notion of a Lie algebra.
There is a close connection between the n-Lie algebras appearing in brane models and strong
homotopy Lie algebras [7]. One can combine the pair of algebras (A , g
A
) into a single space
L. The fundamental identity and the Jacobi identity are then part of a chain of homotopy Jacobi
identities. Hence L is an (ungraded) L
-algebra (or strong homotopy Lie algebra). In this

setting, the generalized Nahm equations are precisely the homotopy MaurerCartan equation for
the L
-algebra L
(R). These L
-structures may play an important role in the quantization of

n-Lie algebras, as we describe later on.
The basic example of a metric n-Lie algebra relevant to the quantization of spheres is denoted
A
n+1
. It is dened as the vector space C
n+1
with basis
1
, . . . ,
n+1
obeying
[
i
1
, . . . ,
i
n
] =
i
1
...i
n
i
n+1
i
n+1
.
This is the unique simple n-Lie algebra over C. Its associated Lie algebra is
g
A
n+1
= so(n+1) .
The problem of constructing associative universal enveloping algebras U (A
n+1
) is addressed in [8]
from the point of view of quantization of duals of n-Lie algebras. In [9], noncommutative Grbner
bases are used to give a monomial basis for U (A
n+1
) realized as a quotient of the free associative
tensor algebra on a basis of A
n+1
.
4
1.3 Geometrical meaning of n-Lie algebras
NambuPoisson manifolds yield geometric realizations of n-Lie algebras by taking A to be the
algebra of functions. They generalize the usual notion of Poisson manifold for n = 2. A Nambu
Poisson structure on a smooth manifold M is an n-Lie algebra structure
{, . . . , } : C
(M)
n
C
(M)
satisfying, in addition to the fundamental identity, the generalized Leibniz rule
{f
1
f
2
, f
3
, . . . , f
n+1
} = f
1
{f
2
, . . . , f
n+1
}+{f
1
, . . . , f
n+1
} f
2
for f
i
C
(M).
The basic example of a NambuPoisson manifold is the unit n-sphere S
n
, regarded as an em-
bedded submanifold of Euclidean space R
n+1
with Cartesian coordinates x
1
, . . . , x
n+1
. The Nambu
Poisson bracket is dened as the extension of the n-bracket
{x
i
1
, . . . , x
i
n
} =
i
1
...i
n
i
n+1
x
i
n+1
by linearity and the generalized Leibniz rule. Thus S
n
provides a geometric realization of the n-Lie
algebra A
n+1
. In particular, the associated Lie algebra g
A
n+1
= so(n+1) generates isometries of
S
n
. Thus we may regard fuzzy n-spheres as the appropriate quantization (to be discussed below) of
either the dual of the n-Lie algebra A
n+1
or of the canonical NambuPoisson structure on S
n
. For
n = 2 both these quantizations are well-known and equivalent.
1.4 Quantum geometry of branes
We have seen that the effective geometry of the D1D3 system in string theory is that of a
fuzzy two-sphere S
2
; this is the most studied and best understood fuzzy space. The question of
a low-energy description of stacks of M2-branes has been the focus of much work in M-theory
over the past few years, and has culminated in the BaggerLambertGustavsson (BLG) theory of
multiple M2-branes [10, 11]. In this article we explore the question of what is the appropriate
notion of noncommutative or deformed geometry that appears in the effective description of M2-
branes and M5-branes, in the context of how the D-brane realizations of monopoles lift to M-
theory. The answers to these questions might provide deep insight into the nature of M-branes
and could signicantly improve the current understanding of M-theory. In particular, we seek a
proper denition of the fuzzy three-sphere S
3
that appears in the M2M5 system. It is expected
that the quantization of the three-sphere will give interesting hints on how to appropriately amend
the current effective descriptions of multiple M2-branes. Moreover, by considering stacks of M2-
branes ending on stacks of M5-branes, the properly dened fuzzy S
3
should also help in pinning
down the gauge structure in an effective description of stacks of multiple M5-branes.
Another context in which theories on quantized spaces arise in low-energy effective descrip-
tions is through modications to generalized Nahm equations when branes are subjected to back-
ground supergravity elds. In [12] it is shown how to interpret the Nahm equations for the D1D3
system as boundary conditions for open strings; they are BPS equations in the effective eld theory
on N coincident D1-branes obtained as the two-dimensional reduction of N = 1 supersymmetric
5
YangMills theory with gauge group U(N) in ten dimensions. In particular, a constant B-eld on
the D3-brane induces a shift in the Nahm equations, which is accounted for by writing a Heisenberg
algebra relation
[X
i
, X
j
] = i
i j

1
on the boundary string elds, with central element
1 and noncommutativity bivector =

1
2
i j
j
a function of the two-form B-eld. The effective noncommutative geometry on the D3-brane
in this case is the well-known Moyal space (see e.g. [13]), which in the present context arises via
quantization of the constant Poisson bracket on R
2
[14, 15, 16, 17].
In an analogous way, the BasuHarvey equations may be regarded as boundary conditions
of open membranes. In this case the modication due to a constant C-eld on the M5-brane is
accounted for by postulating the relations of the NambuHeisenberg algebra [18]
[X
i
, X
j
, X
k
] = i
i jk

1 ,
with 3-central element

1 and noncommutativity 3-vector =
1
3!
i jk
k
a function of the
three-form C-eld. In [12] it is suggested that the associated noncommutative geometry should
arise from quantization of the constant NambuPoisson 3-bracket on R
3
. In the following we will
focus on the geometries mentioned above, and discuss how to make sense of both fuzzy S
3
and the
geometry of the NambuHeisenberg algebra following for a large part the treatment of [8]. These
and other quantized NambuPoisson manifolds have recently been derived as supersymmetric solu-
tions in a 3-Lie algebra reduced model derived from dimensional reduction of the BLG model [19],
which contains the IKKT matrix model as a strong coupling limit.
2. Quantization
2.1 Axioms of quantization
The problem of quantization is highly non-trivial and far from being fully understood. Quan-
tization is essentially a process that turns classical objects into their quantum analogs. At the
classical level, states are points on a Poisson manifold M, and observables are functions on M. At
the quantum level, states are rays in a complex Hilbert space H , and observables are operators on
H . Thus quantization provides a map
C
(M) End(H ) , f

f
that we should subject to a list of axioms depending on the specic applications. We could also
demand that Poisson diffeomorphisms of M map to automorphisms of the associative operator
algebra of quantum observables. If in addition M is a symplectic manifold, then we should further
map Lagrangian submanifolds L M to vectors
L
H .
Diracs wish list was that of a full quantization, which is a map satisfying:
1. The assignment f

f is C-linear, and if f = f
is real then

f =

f
is Hermitian;
2. The identity function f = 1 is mapped to the identity operator id
H
;
3. Correspondence principle: [

f , g] = i h
{f , g} for f , g C
(M); and
6
4. The quantized coordinate functions x
i
act irreducibly on H .
When dealing with homogeneous spaces M = G/H, we may also wish to add:
4a. The Hilbert space H carries a representation of the isometry group G.
The problem with this wish list is that full quantizations in general do not exist; in particular, by
the Groenewold-van Howe theorem there is no full quantization for the symplectic manifolds T
R
n
and S
2
.
There are various loopholes to the obstructions to full quantizations. We shall use Berezin
quantization which is a hybrid of geometric and deformation quantization, and constructs fuzzy
geometry. It utilizes three weakenings of the axioms for quantization. Firstly, one drops the irre-
ducibility assumption, which amounts to a prequantization; this involves a choice of line bundle
over M, which can lead to a quantization of h
1
(i.e. a restriction of the deformation parameter
h to a discrete subset of C
). Secondly, one quantizes only a subset C
(M) of quantiz-
able functions on M; this amounts to choosing a polarization, that corresponds on local charts
to representing M as a cotangent bundle T
U, and leads into geometric quantization. Finally, one

invokes the correspondence principle only to linear order in h; this is the earmark of deformation
quantization.
2.2 Axioms of generalized quantization
In [8] we proposed a generalization of the quantization axioms to NambuPoisson manifolds.
This problem is notoriously difcult, and many previous efforts were devoted to extending geo-
metric quantization (see [8] for a list of references). Here we will extend Berezin quantization to
NambuPoisson manifolds, keeping the data of a complex Hilbert space H and End(H ) as the al-
gebra of quantum observables. The extension requires the formulation of generalized quantization
axioms, that we now describe.
A NambuPoisson quantization is a map Q : End(H ), C
(M), satisfying:
1
. The assignment f

f = Q( f ) is C-linear, and if f = f
is real then

f =

f
is Hermitian;
2
. The identity function f = 1 is mapped to the identity operator id

H
; and
3
. Correspondence principle:
lim
h0
_
_
_
i
h
_
[

f
1
, . . . ,

f
n
]
_
{f
1
, . . . , f
n
}
_
_
_
L
2
= 0
for f
i
C
(M), where : Q() is the symbol map and the L

2
-norm is taken with respect
to a chosen measure on M. If M is a Poisson manifold, then these axioms are always satised for
Berezin quantization.
Axiom 3
requires the choice of an n-Lie bracket on End(H ). A natural choice can often be
obtained by truncating the NambuPoisson algebra on the algebra of polynomials C[x
i
] to obtain a
corresponding n-Lie algebra [20]. Denote the truncated NambuPoisson bracket to polynomials of
degree K by {, . . . , }
K
, and introduce
[

f
1
, . . . ,

f
n
] :=
1
_
i h{(

f
1
), . . . , (

f
n
)}
K
_
7
with K in the classical limit h 0. Then the correspondence principle always holds
automatically. This n-bracket is in general a deformation of the totally antisymmetric operator
product
i
1
i
n
f
i
1

f
i
n
.
3. BerezinToeplitz quantization
We now review the basic aspects of Berezin quantization that we shall need, particularly of
complex projective space. Generally, on a Khler manifold (M, ), we take as Hilbert space the
vector space
H
L
= H
0
(M, L)
of global holomorphic sections of a quantum line bundle L over M with rst Chern class c
1
(L) =
[]; this is the choice of holomorphic polarization. The existence of this holomorphic line bundle
imposes the quantization condition
[] H
2
(M, Z) ,
and the space (M, ) is quantizable only if this constraint is met.
For M = CP
n
, we take the natural Khler two-form corresponding to the FubiniStudy
metric, and set L := O(k). Then the nite-dimensional Hilbert space H
k
:= H
L
consists of
homogeneous polynomials of degree k in the homogeneous coordinates z
0
, z
1
, . . . , z
n
of CP
n
and
can be presented as
H
k
= span
C
(z
1
z
k
)
n
i
=0
= span
C
_
a
1
a
k
|0
_
,
where a
, a
are the standard creation and annihilation operators of an (n +1)-dimensional har-

monic oscillator. This space coincides with the k-particle Hilbert space of lowest Landau level
states in the generalization of the quantum Hall effect (Landau problem) on CP
n
[21].
Generally, an overcomplete basis of the Hilbert space is given by the Rawnsley coherent states.
Associated to any z M there is a corresponding coherent state |z H
L
. For M = CP
n
, the states
|z =
1
k!
_
z
_
k
|0
are the usual Perelomov coherent states. The quantization is now set up by using the coherent states
as a bridge between classical and quantum observables in two ways through the Berezin symbol
and quantization maps
f (z) =
_

f
_
=
z|

f |z
z|z
and

f = Q( f ) =
_
M
n
n!
|zz|
z|z
f .
For M = CP
n
this quantization map denes the fuzzy projective space CP
n
. It obeys the conver-
gence property
lim
k
_
_
_i k
_
Q( f ), Q(g)
Q
_
{f , g}
_
_
_
_
HS
= 0
with respect to the HilbertSchmidt norm on End(H
k
) [22].
8
The Rawnsley coherent states also have geometrical applications. The Bergman metric for
Khler manifolds is given by
g =
1
k
logz|z .
The expansion of this metric in the classical limit k in powers of curvatures approximates
Einstein metrics for projective Khler manifolds, e.g. for CalabiYau manifolds embedded in CP
n
.
To leading orders one has [23]
z|z =
n
+
n1
R
2
+ .
For holomorphically embedded submanifolds M CP
n
, given by the zero locus
M =
_
z CP
n

p(z) = 0
_
of a holomorphic function p(z) in the homogeneous coordinates of CP
n
, one can factor the algebra
of functions on CP
n
by the corresponding ideal and quantize M on the Hilbert space [24]
H
M
=
_
| H
L
p| = 0
_
.
4. Quantization of spheres
2
: The fuzzy sphere
The Berezin quantization of CP
1
is the celebrated fuzzy sphere [25, 26]. For CP
1
= S
2
, the
space of quantizable functions is spanned by spherical harmonics Y
l,m
with l k. The Poisson
bracket
{x
i
, x
j
} =
i jk
x
k
on S
2
maps to the su(2) Lie algebra
[ x
i
, x
j
] = i h
i jk
x
k
under the correspondence principle.
The quantization map is described by the JordanSchwinger transformation which sends the
local coordinates
x
i
=
z
|z|
2
S
2
R
3
to the operators
x
i
=
1
k!
1
a
k1
|00| a
1
a
k1
.
A straightforward calculation gives
h =
2
k
in this case, so that the classical limit is k . This construction generalizes to any projective
space CP
n
by replacing the SU(2) Pauli spin matrices
i
with the Gell-Mann matrices

i
of
its isometry group SU(n+1). The corresponding (non-formal) coherent state star-product f g =
(

f g) is computed explicitly in [27].
9
4
Our quantization of S
4
yields the noncommutative spheres of Guralnik and Ramgoolam [28].
For this, we use the Clifford algebra Cl(R
5
) to isometrically embed S
4
CP
3
via the map
x
i
=
1
|z|
2

i
with x
i
x
i
= 1 .
This embedding is not holomorphic. An alternative quantization in the same spirit is obtained by
considering the sphere bration CP
3
S
2
S
4
[29, 30, 31].
The restricted coherent state projection of the Berezin symbol (

f ),

f H
k
to C
(S
4
)
gives the quantization map
x
i
:=
1
k!
1
a
k1
|00| a
1
a
k1
,
and the endomorphism algebra End(H
k
) consists of the noncommutative polynomials of degree 3k.
This map yields the quadratic Casimir eigenvalue
x
i
x
i
=
_
1+
4
k
_
id
H
k
,
and the 4-Lie bracket is identical to the totally antisymmetric operator product at afne level. In
this case the correspondence principle is satised with the truncated NambuPoisson 4-bracket
{, . . . , }
k
.
3
To quantize the three-sphere, we consider the surjection S
4
S
3
given by x
5
= 0. This
induces an embedding S
3
CP
1
CP
1
CP
3
. However, the constraint x
5
= 0 is not holomorphic,
and so we cannot factorize the operator algebra End(H
k
) by a holomorphic ideal. Thus we must
project the Hilbert space H
k
onto a maximal set of irreducible representations of the isometry
group SO(5) on which the Casimir operator x
i
x
i
is proportional to the identity operator id
H
k
. This
yields a nonassociative operator algebra [28].
The corresponding 3-Lie algebra is dened by [5]
[ x
i
, x
j
, x
k
] := [ x
i
, x
j
, x
k
, x
5
] = i h(k)
i jkl
x
l
.
In previous works, the additional radial fuzziness in the normal directions to S
3
in CP
1
CP
1
CP
3
was dealt with appropriately; either the radial modes are projected out after operator multipli-
cation [32, 28] or are dynamically suppressed [29]. In our case, we choose to keep these modes.
Then the radial fuzziness of our quantum S
3
allows for consistent solutions to the BasuHarvey
equations [33].
4.4 Fuzzy scalar eld theory on S
4
In addition to their relevance in the low-energy descriptions of brane dynamics, quantum eld
theories on Berezin-quantized spaces provide an interesting alternative to lattice regularization.
Let us consider the particular example of the noncommutative four-sphere. The Hilbert space
10
H
k
= H
0
(CP
3
, O(k)) carries both an irreducible representation of the SU(4) isometry group of
CP
3
and the spinor representation of the SO(5) isometry group of S
4
. The Laplace operator
on the algebra of functions C
(S
4
) corresponds to the quadratic Casimir operator of SO(5) in the
spinor representation. It carries over to a linear operator on the endomorphism algebra End(H
k
)
via the Berezin push [34]
B
:= Q .
An action functional for scalar elds may now be constructed through the integral
_
S
4
d
S
4 f =
1
vol(S
2
)
_
CP
3
3
3!
( f ) =
vol(S
4
)
k
tr (

f ) ,
where ( f ) is the image of the function f C
(S
4
) in C
(CP
3
). The corresponding path integral
is dened by taking the integration domain to be the space C
(S
4
) of quantizable functions.
Thus fuzzy quantum eld theories can be used to dynamically distinguish quantum spheres from
quantum projective spaces.
5. Quantization of non-compact manifolds
5.1 Quantization of R
3
The geometry of the quantized NambuHeisenberg algebra
[ x
1
, x
2
, x
3
] = i h

1
can be realized in terms of the noncommutative space R
3
h
which was dened in [35]. In this in-
stance there is no truncated 3-bracket ensuring the correspondence principle. Hence a 3-Lie algebra
structure is only realized at afne level. One possible model for this space is as follows in terms
of R
3
h
.
Take the fuzzy sphere S
2
with Hilbert space
H
k
= H
0
_
CP
1
, O(k)
_
and coordinate generators x
i
obeying [ x
i
, x
j
] =
2i
k

i jk
x
k
. Dene
[ x
1
, x
2
, x
3
] =
i jk
x
i
x
j
x
k
=
6i
k
id
H
k
.
The radius R
k
of this fuzzy sphere, dened by the quadratic Casimir eigenvalue x
i
x
i
= R
2
k
id
H
k
, is
given by
R
k
=
_
1+
2
k
3
_
hk
6
.
The Hilbert space and algebra of quantized functions on R
3
h
are given by
H =
k=1
H
k
and A =
k=1
End(H
k
) .
This describes the space R
3
h
as a discrete foliation of R
3
by fuzzy spheres [35, 36] with radius
R
k
. Note that, as expected from the non-compactness of M in this case, there is no quantization
condition on h C and the Hilbert space H is innite-dimensional.
11
5.2 Quantization of hyperboloids
Our construction can also be extended to the quantization of the hyperboloids H
p,q
, which can
be realized as the non-compact homogeneous space
H
p,q
= SO(p, q)
_
SO(p1, q) ,
or alternatively as the quadric
x
= 1
in R
p+q
with metric (
) of signature (p, q). In this case we must allow for an indenite metric in
the corresponding Clifford algebra. Hence the quantization map involves non-Hermitian operators
and non-unitary representations. Important examples in quantum gravity and string theory involve
fuzzy AdS, and the quantized M5-brane geometry R
1,2
h
R
3
h
. The associated n-Lie algebra A
p,q
is
the unique simple n-Lie algebra over R, where n = p+q1. See [8] for further details.
6. Gerbes and quantization of loop spaces
6.1 Quantization of 2-plectic manifolds
We will now sketch some ways in which one may arrive at a complete quantization of S
3
and
other NambuPoisson manifolds. As we discussed, a symplectic manifold (M, ) is quantizable
if it admits a Hermitian line bundle L with a unitary connection of curvature F
= 2 i .
This two-form is a representative of the rst Chern class c
1
(L) which is a characteristic class of
the line bundle L M, and which is an element of the integer cohomology group H
2
(M, Z).
This class can be constructed explicitly in

Cech cohomology. Given an open cover (U
i
)
iI
of
the manifold M, we construct the curvature two-form F
, the local one-form gauge potentials A

(i)
dening the connection onU
i
, and the transition functions g
(i j)
=g
1
( ji)
: U
i
U
j
U(1) between
neighbouring charts which dene a

Cech one-cocycle, i.e. they obey the cocycle condition
g
(i j)
g
( jk)
g
(ki)
= 1 on U
i
U
j
U
k
.
They are all related through
F
= dA
(i)
on U
i
,
A
(i)
A
( j)
= d logg
(i j)
on U
i
U
j
.
These relationships are obtained directly from repeated applications of Poincars lemma.
The Hilbert space H is then constructed by restricting the space of sections of L. For example,
we saw that in Berezin quantization it is given by the vector space H = H
0
(M, L) of global
holomorphic sections of the quantum line bundle L. A quantization is completed by specifying a
prescription of how to map a certain subset of functions on M to the C
-algebra of linear operators

on H .
Let us attempt to generalize this picture to a 2-plectic manifold (M, ), again endowed with
an open cover (U
i
)
iI
. Such a manifold comes with a closed three-form which gives rise to a
NambuPoisson bracket, analogously to the way in which a symplectic structure gives rise to a
12
Poisson bracket. We thus expect the quantization condition =
i
2
H, where H is a representative
of a class in the integer cohomology group
[] H
3
(M, Z) .
Just like the rst Chern class is a characteristic class for a Hermitian line bundle with connection,
this DixmierDouady class is a characteristic class for a bundle gerbe with connection and curving
H. In the setting of abelian local gerbes this also has a denition in

Cech cohomology. From the
closed three-form H, we obtain by repeated application of Poincars lemma also a 1-connection,
i.e. a family of two-forms B
(i)
on patches U
i
, a 0-connection, i.e. a family of connections
(i j)
=
( ji)
on line bundles L
(i j)
= L
( ji)
over intersections of patches U
i
U
j
representing the rst

Cech
cohomology, and bundle isomorphisms h
(i jk)
: L
(i j)
L
( jk)
L
(ik)
on triple intersections U
i
U
j
U
k
obeying the coherence condition
h
(ikl)
_
h
(i jk)
id
L
(kl)
_
= h
(i jl)
_
id
L
(i j)
h
( jkl)
_
on U
i
U
j
U
k
U
l
.
They are all related through
H = dB
(i)
on U
i
,
B
(i)
B
( j)
= F
(i j)
on U
i
U
j
,
h
(i jk)
(i j)
1+1
( jk)
_
=
(ik)
h
(i jk)
on U
i
U
j
U
k
.
Note that the bundle isomorphisms h
(i jk)
are equivalent to the specication of transition functions
g
(i jk)
: U
i
U
j
U
k
U(1) on triple intersections which dene a

Cech two-cocycle such that
A
(i j)
A
(ik)
+A
( jk)
= d logg
(i jk)
on U
i
U
j
U
k
,
g
(i jk)
g
1
(i jl)
g
(ikl)
g
1
( jkl)
= 1 on U
i
U
j
U
k
U
l
,
where F
(i j)
= dA
(i j)
.
A naive approach to the quantization of a 2-plectic manifold (M, ) would therefore construct
a Hilbert space H from global holomorphic sections of the gerbe, and the C
-algebra corre-
sponding to quantized functions from linear operators on H ; the natural notion of polarization
now appears to be provided by the rst cojet bundle rather than the cotangent bundle [37]. On
M = S
3
= U
1
U
2
with U
1
U
2

= S
2
(1, 1), this may be attained by pullback from the
corresponding structures on S
2
through the Hopf bration : S
3
S
2
; the connection one-form
on this bundle dened via pullback d =
() of the canonical symplectic two-form on S

2
denes a contact structure on S
3
, with volume form d
S
3 = d, and hence in this case one
quantizes the contact manifold (S
3
, ). However, in general all of these objects need to be dened
precisely.
To nd a proper quantization of 2-plectic structures, one can alternatively regard a gerbe as a
sheaf of groupoids with locally isomorphic objects. In this setting one may extend the reinterpre-
tation of geometric quantization in terms of integration of Lie algebroids by Hawkins and others;
see [38] and references therein. A Lie algebroid over M is a vector bundle E M together with
a Lie bracket on the space of sections C
(M, E) and an anchor morphism : E TM to the

13
tangent bundle which implements a Leibniz rule for the bracket; the simplest example is the tangent
Lie algebroid with E = TM and = id
TM
. When M is a point a Lie algebroid is the same thing
as a Lie algebra.
In this setting one starts from the observation that the quantization of the dual of a Lie algebra
g yields a C
-algebra which is the convolution algebra of one of the Lie groups integrating g. This
can be generalized in the following manner. Every quantizable symplectic manifold comes with a
natural Lie algebroid, the Atiyah Lie algebroid which is an extension of the tangent bundle
ad(L) TL
_
U(1)

TM ,
and which can be integrated to a Lie groupoid, abstractly a category with smooth structure in which
every morphism is invertible. The C
-algebra arising in the quantization of symplectic manifolds

should therefore be identied with the convolution algebra of the Atiyah Lie groupoid. The con-
nection on the line bundle L denes a splitting s : TM TL/U(1) of this exact sequence, and
one can construct a Lie algebra homomorphismC
(M) C
(M, TL/U(1)) to the U(1)-invariant

vector elds on the total space of L. In this manner, one can dene polarizations for Lie groupoids,
and reconstruct both geometric and Berezin quantization.
In [39] it was demonstrated that the Lie algebroid of a symplectic manifold is replaced by
the Courant algebroid on a 2-plectic manifold; a Courant algebroid is a vector bundle E M
that generalizes the structure of a Lie algebroid equiped with an inner product on the bres of E.
Given a bundle gerbe with connection, on each patch U
i
one builds the standard Courant algebroid
E
i
:=TU
i
T
U
i
, and then glues together on double intersections using the 0-connection curvatures
F
(i j)
. This gives a Courant algebroid E M which is an extension of the tangent bundle
T
E

TM ,
as well as a splitting s : TM E of this exact sequence given by the 1-connection {B
(i)
}
iI
.
The 2-plectic space (M, ) has a natural L
-algebra structure with de Rham differential d, which

embeds in a canonical L
-algebra associated to the Courant algebroid E M with differential

d
E
=
d. Although the issue of integrating such Courant algebroids has not been completely
settled yet, it is tempting to try to generalize Hawkins approach to quantization to this setting.
One advantage of the groupoid construction, besides its generality, is the fact that it avoids the
construction of a Hilbert space and comes directly to the C
-algebra. For M = S
3
, this algebra
should represent the proper quantization of the enveloping algebra U (A
4
). Given the difculties
with dening the notion of global holomorphic sections of a gerbe, this approach might prove very
fruitful.
6.2 Transgression to loop space
An alternative approach to the quantization of NambuPoisson structures employs a trick to
substitute the gerbe by a principal U(1)-bundle. The price one has to pay is that the base manifold
M is replaced by an innite-dimensional manifold, the loop space of M. Geometric quantization
of loop spaces is discussed in the example M = S
3
in [40], and more generally in the setting of
innite-dimensional Khler geometry in [41]. The appearence of the loop space is very natural in
the context of the quantization of NambuPoisson manifolds. Firstly, this was originally observed
14
in [42] where an action principle for Nambu mechanics on R
3
was formulated as the dynamics of
loops; Nambu mechanics has also been recently proposed as a low energy effective description of
strings and membranes in constant three-form and four-form ux backgrounds [43]. Secondly, the
original attempts [44, 45] to quantize open membranes ending on an M5-brane in a large constant
C-eld background of M-theory proceeded in analogy to the quantization of open strings ending
on a D-brane in a large B-eld [17]. In contrast to the point particle endpoints of open strings,
the boundaries of open membranes are closed strings, and this leads to a noncommutative loop
space structure. In [44, 45] canonical quantization of the M5-brane theory was considered with the
natural Poisson structure on loop space, leading to a complicated nonassociative algebraic structure.
Finally, the map to the loop spaces of R
4
and S
3
has proved to very helpful in generalizations of
the ADHMN construction of monopoles (D1-branes) to the self-dual strings of the six-dimensional
theory on an M5-brane [46, 47].
This trick is called a transgression and works as follows. Consider the correspondence
LMS
1
ev
zzu
u
u
u
u
u
u
u
u
_
S
1
%%
K
K
K
K
K
K
K
K
K
K
M LM
between M and its free loop space LM. Here ev is the obvious evaluation map of the loop at the
given angle in S
1
, and
_
S
1 is the integral over the angle parameterizing the loop. The transgression
map T :
+1
(M)
(LM) amounts to the pullback along ev and the pushforward along

_
S
1 ,
i.e. T =
__
S
1
_
!
ev
. Explicitly, it is given by
(T )
x
_
v
1
(), . . . , v
k
()
_
=
_
S
1
d
_
v
1
(), . . . , v
k
(), x()
_
,
k+1
(M)
for x M, where we used the natural tangent vector x() LTM = TLM available on loop
space to ll one of the slots of the (k +1)-form on M.
Using T , we can thus map the DixmierDouady class on a 2-plectic manifold M to a rst
Chern class on the free loop space LM. Put differently, a bundle gerbe on M gives rise to a line
bundle on LM. In principle, the latter can then be quantized in the usual way. The difculty here
is that one is working with the innite-dimensional base space LM.
Nevertheless, the transgression map can be used to connect the different perspectives on the
quantum geometry of M5-branes. The conventional approach, cf. e.g. [12], yields a noncommuta-
tive space with the description
_
x
i
, x
j
, x
k
= i
i jk

1 ,
where [, , ] is some 3-algebraic structure, e.g. the bracket of a 3-Lie algebra. Its transgression
corresponds to noncommutative loop space relations
_
x
i
(), x
j
()
= i
i jk
x
k
() ( )
1 .
This coincides with the noncommutative loop space structure derived in [44, 45].
15
Acknowledgments
RJS would like to thank Dorothea Bahns, Harald Grosse and George Zoupanos for the in-
vitation and kind hospitality at the Corfu Summer Institute. This work was supported by grant
ST/G000514/1 String Theory Scotland from the UK Science and Technology Facilities Council.
The work of CS was supported by a Career Acceleration Fellowship from the UK Engineering and
Physical Sciences Research Council.
References
[1] W. Nahm, A simple formalism for the BPS monopole, Phys. Lett. B 90 (1980) 413414.
[2] D. -E. Diaconescu, D-branes, monopoles and Nahm equations, Nucl. Phys. B 503 (1997) 220238
[hep-th/9608163].
[3] D. Tsimpis, Nahm equations and boundary conditions, Phys. Lett. B 433 (1998) 287290
[hep-th/9804081].
[4] R. C. Myers, Dielectric branes, JHEP 12 (1999) 022 [hep-th/9910053].
[5] A. Basu, J. A. Harvey, The M2M5 brane system and a generalized Nahms equation, Nucl. Phys. B
713 (2005) 136150 [hep-th/0412310].
[6] V. T. Filippov, n-Lie algebras, Sib. Mat. Zh. 26 (1985) 126140.
[7] C. Iuliu-Lazaroiu, D. McNamee, C. Saemann, A. Zejak, Strong homotopy Lie algebras, generalized
Nahm equations and multiple M2-branes, 0901.3905 [hep-th].
[8] J. DeBellis, C. Saemann, R. J. Szabo, Quantized NambuPoisson manifolds and n-Lie algebras, J.
Math. Phys. 51 (2010) 122303 [1001.3275 [hep-th]].
[9] M. R. Bremner, H. A. Elgendy, Universal associative envelopes of (n +1)-dimensional n-Lie
algebras, 1008.1987 [math.RA].
[10] J. Bagger and N. Lambert, Gauge symmetry and supersymmetry of multiple M2-branes, Phys. Rev. D
77 (2008) 065008 [0711.0955 [hep-th]].
[11] A. Gustavsson, Algebraic structures on parallel M2-branes, Nucl. Phys. B 811 (2009) 6676
[0709.1260 [hep-th]].
[12] C. -S. Chu, D. J. Smith, Towards the quantum geometry of the M5-brane in a constant C-eld from
multiple membranes, JHEP 04 (2009) 097 [0901.1847 [hep-th]].
[13] R. J. Szabo, Quantum eld theory on noncommutative spaces, Phys. Rept. 378 (2003) 207299
[hep-th/0109162].
[14] F. Ardalan, H. Arfaei, M. M. Sheikh-Jabbari, Noncommutative geometry from strings and branes,
JHEP 02 (1999) 016 [hep-th/9810072].
[15] C. -S. Chu, P. -M. Ho, Noncommutative open string and D-brane, Nucl. Phys. B 550 (1999) 151168
[hep-th/9812219].
[16] V. Schomerus, D-branes and deformation quantization, JHEP 06 (1999) 030 [hep-th/9903205].
[17] N. Seiberg, E. Witten, String theory and noncommutative geometry, JHEP 09 (1999) 032
[hep-th/9908142].
16
[18] Y. Nambu, Generalized Hamiltonian dynamics, Phys. Rev. D 7 (1973) 24052412.
[19] J. DeBellis, C. Saemann, R. J. Szabo, Quantized NambuPoisson manifolds in a 3-Lie algebra
reduced model, 1012.2236 [hep-th].
[20] P. -M. Ho, R. -C. Hou, Y. Matsuo, Lie 3-algebra and multiple M2-branes, JHEP 06 (2008) 020
[0804.2110 [hep-th]].
[21] D. Karabali, V. P. Nair, Quantum Hall effect in higher dimensions, matrix models and fuzzy geometry,
J. Phys. A 39 (2006) 1273512764 [hep-th/0606161].
[22] M. Bordemann, E. Meinrenken, M. Schlichenmaier, Toeplitz quantization of Khler manifolds and
gl(N), N limits, Commun. Math. Phys. 165 (1994) 281296 [hep-th/9309134].
[23] S. Zelditch, Szeg o kernels and a theorem of Tian, Int. Math. Res. Notices 6 (1998) 317331
[math-ph/0002009].
[24] C. Saemann, Fuzzy toric geometries, JHEP 02 (2008) 111 [hep-th/0612173].
[25] F. A. Berezin, General concept of quantization, Commun. Math. Phys. 40 (1975) 153174.
[26] J. Madore, The fuzzy sphere, Class. Quant. Grav. 9 (1992) 6988.
[27] A. P. Balachandran, B. P. Dolan, J. -H. Lee, X. Martin, D. OConnor, Fuzzy complex projective spaces
and their star products, J. Geom. Phys. 43 (2002) 184204 [hep-th/0107099].
[28] Z. Guralnik, S. Ramgoolam, On the polarization of unstable D0-branes into noncommutative odd
spheres, JHEP 02 (2001) 032 [hep-th/0101001].
[29] J. Medina, D. OConnor, Scalar eld theory on fuzzy S
4
, JHEP 11 (2003) 051 [hep-th/0212170].
[30] B. P. Dolan, D. OConnor, A fuzzy three-sphere and fuzzy tori, JHEP 10 (2003) 060
[hep-th/0306231].
[31] Y. Abe, A construction of fuzzy S
4
, Phys. Rev. D 70 (2004) 126004 [hep-th/0406135].
[32] H. Grosse, C. Klimcik, P. Presnajder, Finite quantum eld theory in noncommutative geometry,
Commun. Math. Phys. 180 (1996) 429438 [hep-th/9602115].
[33] H. Nastase, C. Papageorgakis, S. Ramgoolam, The fuzzy S
2
structure of M2M5 systems in ABJM
membrane theories, JHEP 05 (2009) 123 [0903.3966 [hep-th]].
[34] C. Iuliu-Lazaroiu, D. McNamee, C. Saemann, Generalized Berezin quantization, Bergman metrics
and fuzzy Laplacians, JHEP 09 (2008) 059 [0804.4555 [hep-th]].
[35] A. B. Hammou, M. Lagraa, M. M. Sheikh-Jabbari, Coherent state induced star product on R
3
and the
fuzzy sphere, Phys. Rev. D 66 (2002) 025025 [hep-th/0110291].
[36] E. Batista, S. Majid, Noncommutative geometry of angular momentum space U (su(2)), J. Math.
Phys. 44 (2003) 107137 [hep-th/0205128].
[37] J. C. Baez, A. E. Hoffnung, C. L. Rogers, Categoried symplectic geometry and the classical string,
Commun. Math. Phys. 293 (2010) 701725 [0808.0246 [math-ph]].
[38] E. Hawkins, A groupoid approach to quantization, J. Sympl. Geom. 6 (2008) 61125
[math.SG/0612363].
[39] C. L. Rogers, 2-plectic geometry, Courant algebroids, and categoried prequantization, 1009.2975
[math-ph].
17
[40] J. L. Brylinski, Loop Spaces, Characteristic Classes and Geometric Quantization, Birkhuser,
Boston, 1993.
[41] A. Sergeev, Khler Geometry of Loop Spaces, World Scientic, Singapore, 2008.
[42] L. Takhtajan, On foundation of the generalized Nambu mechanics (second version), Commun. Math.
Phys. 160 (1994) 295316 [hep-th/9301111].
[43] C. -S. Chu, P. -M. Ho, D1-brane in constant RR 3-form ux and Nambu dynamics in string theory,
1011.3765 [hep-th].
[44] E. Bergshoeff, D. S. Berman, J. P. van der Schaar, P. Sundell, A noncommutative M-theory ve-brane,
Nucl. Phys. B 590 (2000) 173197 [hep-th/0005026].
[45] S. Kawamoto, N. Sasakura, Open membranes in a constant C-eld background and noncommutative
boundary strings, JHEP 07 (2000) 014 [hep-th/0005123].
[46] A. Gustavsson, Self-dual strings and loop space Nahm equations, JHEP 04 (2008) 083 [0802.3456
[hep-th]].
[47] C. Saemann, Constructing self-dual strings, 1007.3301 [hep-th].
18

1101.5987v1 Saemann 2

Uploaded by

Copyright:

Available Formats

1101.5987v1 Saemann 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1101.5987v1 Saemann 2

Uploaded by

Copyright:

Available Formats

a

Report numbers: HWM111 , EMPG1103

-algebra (or strong homotopy Lie algebra). In this

-structures may play an important role in the quantization of

1 and noncommutativity bivector =

). Secondly, one quantizes only a subset C

U, and leads into geometric quantization. Finally, one

. The identity function f = 1 is mapped to the identity operator id

(M), where : Q() is the symbol map and the L

are the standard creation and annihilation operators of an (n +1)-dimensional har-

with the Gell-Mann matrices

, the local one-form gauge potentials A

-algebra of linear operators

() of the canonical symplectic two-form on S

(M, E) and an anchor morphism : E TM to the

-algebra arising in the quantization of symplectic manifolds

(M, TL/U(1)) to the U(1)-invariant

-algebra structure with de Rham differential d, which

-algebra associated to the Courant algebroid E M with differential

(LM) amounts to the pullback along ev and the pushforward along

You might also like