From the Maxwell’s equations to the String Theory: new possible mathematical

connections

Michele Nardelli1,2, Rosario Turco

Dipartimento di Scienze della Terra

1

Università degli Studi di Napoli Federico II, Largo S. Marcellino, 10

80138 Napoli, Italy

Dipartimento di Matematica ed Applicazioni “R. Caccioppoli”

2

Università degli Studi di Napoli “Federico II” – Polo delle Scienze e delle Tecnologie
Monte S. Angelo, Via Cintia (Fuorigrotta), 80126 Napoli, Italy

Abstract

In this paper in the Section 1, we describe the possible mathematics concerning the unification
between the Maxwell's equations and the gravitational equations. In this Section we have described
also some equations concerning the gravitomagnetic and gravitoelectric fields. In the Section 2, we
have described the mathematics concerning the Maxwell's equations in higher dimension (thence
Kaluza-Klein compactification and relative connections with string theory and Palumbo-Nardelli
model). In the Section 3, we have described some equations concerning the noncommutativity in
String Theory, principally the Dirac-Born-Infeld action, noncommutative open string actions,
Chern-Simons couplings on the brane, D-brane actions and the connections with the Maxwell
electrodynamics, Maxwell's equations, B-field and gauge fields. In the Section 4, we have described
some equations concerning the noncommutative quantum mechanics regarding the particle in a
constant field and the noncommutative classical dynamics related to quadratic Lagrangians
(Hamiltonians) connected with some equations concerning the Section 3. In conclusion, in the
Section 5, we have described the possible mathematical connections between various equations
concerning the arguments above mentioned, some links with some aspects of Number Theory
(Ramanujan modular equations connected with the phisycal vibrations of the superstrings, various
relationships and links concerning π, φ thence the Aurea ratio), the zeta strings and the Palumbo-
Nardelli model that link bosonic and fermionic strings.

1
 Introduction

Maxwell's Equations are an important connection between

mathematics and physics, an instrument without which we would
not have reached the present state of knowledge such as relativity,
quantum mechanics and string theory.

These equations are of extraordinary importance, especially from a

physical point of view, as well as mathematics.
The authors, in this paper, show how to treat them from a
mathematical point of view with techniques of vector calculus.
James Clerk Maxwell

In the second part the authors examine, with the method of the conceptual experiments and the
"umbral calculus", a "symmetrization" of these equations and we have the question of the
correctness of the approach and the existence of phenomena that should be linked then this fact.
All this is done without introducing the Minkowski space and the equations of Einstein-Maxwell,
but keeping for simplicity only on Maxwell's equations.

Another basic question that accompanies the whole paper is the possibility or not to create anti-
gravity and propulsion systems, in this regard we have mentioned the EHT theory, a theory to
unify relativity and quantum mechanics, that is multi-dimensional (8 dimensions) and that can be
related with M-theory and superstring theory.

2
1. On the possibile mathematics concerning the unification between the
Maxwell’s equations and the gravitational equations.

Maxwell's equations

Maxwell's equations are:

Law Equation
Gauss's law of electric field r ρ
∇•E = (1)
ε0
r
Gauss's law of magnetic field ∇ • B = 0 (2)
r
Faraday’s law on the electric field r ∂B
∇× E = − (3)
∂t
r
Ampere’s law on the magnetic field r r ∂E
∇ × B = µ 0 J + µ 0ε 0 (4)
∂t

where:
∇ it’s is the mathematical symbol "nabla" (gradient)
∇• it’s the mathematical symbol for the divergence
∇ × it’s is the mathematical symbol of the rotor

Then:
r
E the electric field
r
B the magnetic induction
r
H the magnetic field
r
J the current density
ε0 the dielectric constant of vacuum 8,854188 x 10-12 F/m
µ0 the magnetic permeability of vacuum 4π x 10-7 H/m
co the speed of light 3 x 108 m/s

Some useful rule for the follow:

( )
r
• The divergence of a rotor is always zero or ∇ • ∇ × C = 0
• The rotor of a gradient is always zero or ∇ × (∇C ) = 0
• The divergence of a gradient is the Laplacian or ∇ • (∇C ) = ∇ 2 C
( ) ( )
r r r
• The rotor of a rotor instead is ∇ × ∇ × C = ∇ ∇ • C − ∇ 2 C

3
Maxwell's equations in vacuum
We start from equation (4). We divide both sides by µ0 and bearing in mind that:
1
= µ 0ε 0
c02

where c0 is the speed of light, we obtain:

r r
r J ∂E
c02 ∇ × B = + (4’)
ε 0 ∂t
Recall that the Lorentz’s force is given by:
r
( r r r
F =q E+v×B ) (5)

The equations, including (4 ') and (5), are the equations in a local form and permit the calculation
of electromagnetic fields in vacuum from known values of charge density ρ and current density J.

Maxwell's equations in materials

If we are inside of the materials then the electromagnetic waves must take account of other
physical phenomena due to the electric induction D, the polarization P, the magnetic induction M
such that:
r r r
D = ε0E + P
r r r
(
B = µ0 H + M )
P and M are the mean value of electric and magnetic dipole per unit volume.

If we consider the medium as linear and isotropic material, the equations are simplified in:
r r
D = ε 0ε r E
r r
B = µ0 µr H (6)
µ = µ0 µ r
ε = ε 0ε r

Thence, we obtain:

Law Equation
r
Gauss's law of electric field ∇•D = ρ (1’)
r
Gauss's law of magnetic field ∇•B = 0 (2)
r
Faraday’s law on the electric field r ∂B
∇× E = − (3)
∂t
r
Ampere’s law on the magnetic field r r ∂D
∇× H = J + (4’’)
∂t

where

4
µr is the relative magnetic permeability

The eq. (3) can lead to something more interesting through (6) and the precedent rules :
r
r ∂B
∇× E = −
∂t
r r r
( r
)
∇ × ∇ × E = −µ 0 µ r
∂
∂t
( r
) ∂  r ∂D 
∇ × H = − µ  J +
∂t 
∂r
 = −µ  J + ε
∂t  ∂t 
∂E  ∂2E
 = − µε 2
∂t  ∂t

thence, we obtain:
r
r ∂2E
∇ E = − µε 2
2
(7)
∂t
1
In the vacuum ρ=0, J=0 and µε = 2 with v that is the speed of light , then we obtain that:
v r
2
r r
( ) r
(
∇ E = ∇ ∇ • E −∇× ∇× E = 2 2
1 ∂2E
v ∂t
) (8)

which is the D'Alembert wave equation in vacuum (v = c0). For the vector B we can use a similar
procedure.

Vector potential and scalar potential

If in (2) the divergence is zero using the fact that the divergence of a rotor always gives a null
result, then there exists a vector potential A, such that:
r r
B = ∇× A
r
∇•B = 0 (9)

From (3) thence, we have that:

r ∂ r
∇× E = − ∇× A
∂t
r ∂ r
∇× E + ∇× A = 0
∂t
 r ∂ r
∇ ×  E + A = 0
 ∂t 

With the last expression, using the rule that the rotor of a gradient is always zero, then,
introducing a scalar potential φ is:
r ∂ r
E + A = −∇φ
∂t
thence, we have that:
r  ∂ r
E = − ∇φ + A  (10)
 ∂t 

5
Utilizing the eqs. (9),(10) we can rewrite the equations (1),(4) as follows:

r  ∂ r  ∂ r ρ
∇ • E = −∇ •  ∇φ + A  = − ∇ 2φ + A  =
 ∂t   ∂t  ε 0
It’s equivalent to:
∂ r ρ
∇ 2φ + A=− (11)
∂t ε0
From (4), instead, is:
r r 1 ∂ ∂ r
∇ × ∇ × A = µ 0 J − 2  ∇φ + A 
c ∂t  ∂t 
r 0
( )
r J ∂
c02 ∇ × ∇ × A = −  ∇φ + A 
ε 0 ∂t 
∂ r
∂t 
r r
r 2
( )
r ∂ ∂2 A J
− c 0 ∇ A + c 0 ∇ ∇ • A + ∇φ + 2 =
2 2

∂t ∂t ε0
(12)

Gauge transformations
Let's see what happens if on the vector potential A and scalar potential φ are made changes like:
ur ur
A → A + ∇ψ
∂ψ
φ →φ −
∂t
Considering that the divergence of a rotor is null and that the rotor of a gradient is zero, then the
equations of vectors B and E do not change (see (9) and (10)). So we speak of gauge invariance.

We can use gauge invariance and choose the vector A as appropriate, for example:

r 1 ∂φ
∇• A= − 2
c0 ∂t
Now with (1) and (10) we obtain that:

r  ∂ r
E = − ∇ φ + A 
 ∂t 
r  ∂ r ∂ r 1 ∂ 2φ ρ
∇ • E = −∇ •  ∇φ + A  = −∇ 2φ − ∇A = −∇ 2φ + 2 2 =
 ∂t  ∂t c0 ∂t ε

so
1 ∂ 2φ ρ
∇ 2φ − =− (13)
c0 ∂t
2 2
ε
If we substitute in (12) we obtain:
ur
ur ∂ 2 A ur
∇ A − 2 = − µ0 J
2
(14)
∂t

6
Now (13) and (14) constitute a system of 4 equations or a four-vector that describes the waves
advancing in space-time with speed c0. In fact in (14) the vectors A and J can be decomposed into
the components in the direction x, y, z.

In fact, you solve the Maxwell equations by introducing a vector potential and a scalar potential,
then gauge invariance is exploited by reducing everything to a system of differential equations in
four scalar functions as follows:

1 ∂ 2φ ρ
∇ φ− 2 2 =−
2

c0 ∂t ε
∂ 2 Ax
∇ 2 Ax − = − µ0 J x
∂t 2
∂ 2 Ay
∇ Ay − 2 = − µ0 J y
2

∂t
∂2 A
∇ 2 Az − 2 z = − µ0 J z
∂t

These equations put in evidence that the behavior of electromagnetic waves and of light were
aspects of the same phenomenon. So far the classical theory.

With regard the problem of symmetrization of the equations.

Following Heim and Hauser (see [1]), we start our simplistic flight of fancy, a little as Maxwell,
who discovered from a mathematical point of view that missing something to the Ampere's law
and introduced the temporal variation of electric induction D.

The aim is to see if we can obtain the symmetrization of equations, adding the terms used and
that in practice are negligible, and introduce a formulation of linearized gravity, without
necessarily introducing the Minkowski space and the equations of Einstein-Maxwell, but starting
simply by classical equations of Maxwell.

If we look at (1 ') (2) (3) (4''), described below for convenience, we note that, in "a sense", these are
two by two similar, but between (1') and (2) there is an obvious "symmetry breaking", this at least
from a formal mathematical point of view:
r
∇•D = ρ (1’)
r
∇•B = 0 (2)
ur
ur ∂B
∇× E = −
∂t (3)
ur
uur ur ∂ D
∇× H = J + (4’’)
∂t

In (2), for example, there isn’t a "density of magnetic charge" ρm as in (1 '), while in (3) compared
to (4'') there isn’t a corresponding "density of magnetic current" Jm.

7
Furthermore between (3) and (4'') are opposite also the signs of the partial derivatives.

In the reality should be borne in mind that a magnetic charge is non-existent, but in a fantasy of
symmetrization, especially for a conceptual experiment and "umbral calculus", leads us to say
that the equations (2) and (3), which are true compared with experiments, are such because the
boundary conditions involving the negligible or null terms as "density of magnetic charge" and
"density of magnetic current".

We start writing a first form of equations of symmetrization:

r
∇•D = ρ (1’)
r
∇ • B = ρm (2a)
ur
ur uur ∂ B
∇ × E = Jm − (3a)
∂t
ur
uur ur ∂ D
∇× H = J + (4’’)
∂t

where in the (2a) (3a) we have introduced the density missing.

Can we add the gravity strength to the Maxwell's equations?

We begin another flight of fancy: it is possible to envisage the introduction of gravity in these
equations? Or at least an effect of fields that interacts with it?

We know that there exist an analogy between the Coulomb's law and the Gravity’s law; the
Coulomb's law expresses the interaction between two electric charges q1 and q2:

1 q1 ⋅ q2
F=
4πε 0 r 2

The strength of gravity expresses the interaction between two masses m1 and m2:

m1 ⋅ m2
F =G
r2
where G = 6,670 x 10-11 Nm2Kg-2

In correspondence of the Coulomb’s strength we have an electric field expressible by:

1 q
E=
4πε 0 r 2

The gravitational field can be expressed with:

m
Xg = G (15)
r2

8
Can we create an electric field equivalent to a gravitational field? If we equate the two fields we
see that we must introduce a "charge equivalent to the mass":

1
q = G ⋅ m → qeq = 4πε 0G ⋅ m
4πε 0
so:

qeq = k g m
(16)
k g = 4πε 0G

g= 6.67422 x 10-11 m³/s² kg

We would also include a charge density equivalent to the mass ρem through a mass density per
unit volume ρmv:

ρ em = k g ρ mv (17)

Furthermore, if we want to write, with regard the gravity, something of equivalent to (1 ') we must
to consider a parameter η0 = ε0 such that:

kg
η0 = (18)
4π G

The (15), then, could be written as follows:

1 qeq
Xg = (19)
4πη0 r 2
ur ur
At this point for the gravity you need to locate an equivalent of D = ε 0 E (for linear and isotopic
ur uuur
materials), that is R = η0 X g ; thence the equivalent of (1') for the gravity would be:

r
∇ • R = k g ρ mv = ρ em (20)
while:
uur
∇× X = 0 (21)

The (21) is a result of the fact that the gravitational field should be irrotational or conservative,
but we will see that we have to work again on this up and perhaps discover news unexpected.

Physical correctness of the idea.

Is it correct from the physical point of view, what we have obtained from a formal point of view?
In practice, because of the similarity of the fields, it was said that you could get an electric field
equivalent to gravity (at least in terms of strength).

9
What are the differences? In the Coulomb strength we can have both positive and negative
electric charges and strengths of attraction if the charges are of a different sign, repulsive if the
charges are of equal sign, while in the gravitational strength we have only positive mass and the
gravitational strength is only attractive.

However, isn’t the direction of the force, which reduces or invalidates the idea, the direction must
be different, but it is certainly true that "a flow of charges is also a flow of masses", so there are
electric currents and current of masses, the current of masses are in this case both sources of
gravity field and of electric and magnetic field. So the only valid physical sense here is that we are
introducing fields equivalent to gravitational fields, or gravito-electric and gravito-magnetic fields.

We try to unify the equations [1b]

Now suppose that we could put together all the equations are those of Maxwell, or to obtain an
unification of electromagnetic and gravitational equations ensuring the symmetrization and the
signs. How to proceed with a hypothetical unification of equations?

It involves:

• to introduce the terms of connection:

the “current density of mass displacement”, represented by the partial derivative in
time of R;

the “current density mass” Jk;
• to introduce a constant γ0 for the dimensional consistency of the physical dimensions
• maintain a symmetry of the equations and signs (X should have a formal appearance as
similar to E ¸ E and H must have, with opposite signs, contributions of link because the
behaviour in the classical Maxwell's equations)

We get a unified system of six equations:

r
∇•D = ρ (1’)
r
∇ • B = ρm (2a)
r
∇ • R = k g ρ mv = ρ em (2b)
ur ur
uur uur ∂ B ur ∂ D
∇ × X = kg ( J m + ) − kgγ 0 ( J + ) (2c)
∂urt ur ∂ t
ur uur ∂ R uuuur ∂ B
∇ × E = γ 0 (Jk + )− J m − (3a)
∂t ∂t
ur ur
uur ur ∂ D uur ∂ R
∇× H = J + − Jk − (4’’’)
∂t ∂t

1 µ0
γ0 = =
4πε 0 c0 4π c0

10
The verification of laws [1b]

After the formal-mathematical symmetrization and unification we must verify that the equations
do not violate the laws of conservation of charge and irrotational fields. The previous equations,
ur
∂R
in the absence of currents of mass and time-varying gravity field, will be such that Jk=0 and =
∂t
0; the variations are due to movement of mass. An electric current is either a flow of charge but
also of mass.

Conservation of total charge

Take the (4'' ') and calculate the divergence of a rotor (that is null). We obtain that:
r r
( r
)  r ∂D r ∂R 
∇ • ∇ × H = ∇ •  J +
∂t
− Jk −
∂t 
r ∂
∂t
r r ∂
∂t
r
 = ∇ • J + ∇ • D − ∇ • J k − ∇ • R = 0


From (1’) (2b), we have that:

r ∂ρ r ∂ρ
∇•J + − ∇ • J k − em = 0
∂t ∂t

If we pass to the integral of surface, obtain the law of conservation of charge:

∂Qem ∂Q
I − Ik − = − (22)
∂t ∂t
ur
∂R
It 's clear that in cases where Jk=0 and = 0 we get the actual law of conservation of charge
∂t
∂Q
I =− that we obtain from (4’’), and that represents an equation of continuity.
∂t
In the event that we wanted to consider all the terms of (22) we observe that the flow of electrons
is made up of electronic mass m = 9,1091 x 10-31 Kg, while the constant 4πε0G is 1,8544 x 10-21
Kg-3 m-1s2C2N for which the mass flow of electrons is of the order of 10-51. If we assume the
presence only of the magnetostatic field, therefore only no time-varying magnetic field and in the
∂Q
absence of electric charges, then only Jk is different from zero and I k = − em ; but also here it is
∂t
negligible, because taking 4πε0G, a mass of 1 kg per second causes a current 6.18 x 10-21 A.

Apparent irrotational electric field?

In particular, the (3a) would seem to violate the irrotational electric field, saying that one has an
electric field perpendicular to the mass movement and arranged in a circle around the mass. In
fact the irrotational electric field in a magnetostatic field is valid, this is because that the mass
flow is significant is need to achieve enormous amounts like 1012 tons.

New phenomenon?

11
The (2c) says that in the presence of non static electric and magnetic fields, then the gravitational
field is not irrotational, taking into account that there is no a magnetic current density (Jm = 0). In
fact from what has so far suggest to us the formulas and calculations, the mass flows very little
influence on electromagnetic fields, but it is certainly possible otherwise. An idea, therefore, is to
exploit the (2c) to create the opposite strengths to the gravitation, or to obtain an
electromagnetic levitation (not just the antigravity! But we are very close like effect).

1.1 On some equations concerning the Gravitoelectric and Gravitomagnetic fields [1]

Considering the Einstein-Maxwell formulation of linearized gravity, a remarkable similarity to the

mathematical form of the electromagnetic Maxwell equations can be found. In analogy to
electromagnetism there exist a gravitational scalar and vector potential, denoted by Φ g and Ag ,
respectively. Introducing the gravitoelectric and gravitomagnetic fields

e = −∇Φ g and b = ∇ × Ag (23)

the gravitational Maxwell equations can be written in the following form:

16πG
∇ ⋅ e = −4πGρ , ∇⋅b = 0 , ∇×e = 0 , ∇×b = − j (24)
c2

where j = ρv is the mass flux and G is the gravitational constant. The field e describes the
gravitational field form a stationary mass distribution, whereas b describes an extra gravitational
field produced by moving masses.
At critical temperature Tc some materials become superconductors that is, their resistance goes
to 0. Superconductors have an energy gap of some E g ≈ 3.5kTc . This energy gap separates
superconducting electrons below from normal electrons above the gap. At temperatures below
Tc , electrons (that are fermions) are coupled in pairs, called Cooper pairs, which are bosons.
We note that, in this case, there is the application of the relationship of Palumbo-Nardelli model,
i.e.

− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
− ∫ d 26 x g −
 16πG 8 2 

( )
∞
 1 ~ 2 κ 2
2 
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  ,
1 1/ 2
(25)
0
2κ 10  2 g10 

a general relationship that links bosonic and fermionic strings acting in all natural systems.
A rotating superconductor generates a magnetic induction field, the so called London moment
2m
B = − e ω (26)
e

where ω is the angular velocity of the rotating ring and e denotes elementary charge. It should
be noted that the magnetic field in Tajmar’s experiment [1] is produced by the rotation of the
ring, and not by a current of Cooper pairs that are moving within the ring.
Tajmar and his colleagues simply postulate an equivalence between the generated B field, eq.
(26) with a gravitational field by proposing a so called gravitomagnetic London effect.

12
Let R denote the radius of the rotating ring, then eq. (26) puts a limit on the maximal allowable
magnetic induction, Bmax , which is given by

me k BTC
2
Bmax = 14 . (27)
e2 R 2

If the magnetic induction exceeds this value, the kinetic energy of the Cooper pairs exceeds the
maximum energy gap, and the Cooper pairs are destroyed. The rotating ring is no longer a
superconductor. Moreover, the magnetic induction must not exceed the critical value Bc (T ) ,
which is the maximal magnetic induction that can be sustained at temperature T , and is
dependent on the material. In this possible scenario, the magnetic induction field B is equivalent
to a gravitophoton (gravitational) field bgp . Therefore, the following relation holds, provided that
B is smaller than Bmax

B
bgp ∝ B . (28)
Bmax

As soon as B exceeds Bmax the gravitophoton field vanishes. Can be derived the following
general relationship between a magnetic and the neutral gravitophoton field, bgp :

 1  e B
bgp =  − 1 B (29)
 (1 − k )(1 − ka )  me Bmax

where k = 1 / 24 and a = 1 / 8 . The dimension of bgp is s −1 . Inserting eq. (26) into eq. (29), using
eq. (28), and differentiating with respect to time, results in

∂bgp  1  2e B ∂B
= − 1 . (30)
∂t  (1 − k )(1 − ka )  me Bmax ∂t

Integrating over an arbitrary area A yields

∂bgp
∫ ∂t
⋅ dA = ∫ g gp ⋅ ds (31)

where it was assumed that the gravitophoton field, since it is a gravitational field, can be
separated according to eqs. (23), (24). Combining eqs. (30) and (31) gives the following
relationship

 
 1  2e B ∂B
∫ gp
g ⋅ ds =  − 1 ∫
 1 − 1 1 − 1   me Bmax ∂t
⋅ dA . (32)
  24  24 ⋅ 8  

From eq. (26) one obtains

13
 ∂B 2m
= − e ω& . (33)
∂t e

Next, we apply eqs. (32) and (33) calculating the gravitophoton acceleration for the in-ring
accelerometer. It is assumed that the accelerometer is located at distance r from the origin of
the coordinate system. From eq. (26) it can be directly seen that the magnetic induction has a z -
component only. From eq. (32) it is obvious that the gravitophoton acceleration is in the r - θ
plane. Because of symmetry reasons the gravitophoton acceleration is independent on the
azimuthal angle θ , and thus only has a component in the circumferential (tangential) direction,
denoted by êθ . Since the gravitophoton acceleration is constant along a circle with radius r ,
integration is over the area A = πr 2eˆz . Inserting eq. (33) into eq. (32), and carrying out the
integration the following expression for the gravitophoton acceleration is obtained

1 B
g gp = − ω& r (34)
10 Bmax

where the minus sign indicates an acceleration opposite to the original one and it was assumed
that the B fields is homogeneous over the integration area.
The ratio of the magnetic fields was calculated from the following formula, obtained by dividing
eq. (26) by the square root of eq. (27)

B 1  me 
=  ωR . (35)
Bmax 7  k BTC 

Inserting an estimated average value of ω = 175 rad/s, me = 9 × 10 −31 kg, k B = 1.38 × 10 −23 J/K,
TC = 9.4 K, and R = 7.2 × 10 −2 m, this ratio is calculated as 3.97 × 10 −4 .

2. On the mathematics concerning the Maxwell’s equations in higher dimensions [2]

In (3+1)-D, the field strength tensor F can be represented as an anti-symmetric 4 × 4 matrix:

 0 F 01 F 02 F 03 
 
 − F 01 0 F 12 F 13 
F=  . (36)
− F
02
− F 12 0 F 23 
 − F 03 − F 13 − F 23 0 


{ } { }
We define the controvariant spacetime position vector as x µ ≡ x 0 ≡ ct , x1 , x 2 , x 3 , where the
lowercase Greek letters represent spacetime indices {µ ,ν } = {0,1,2,3} . We choose the metric
tensor g = diag {− 1,1,1,1} to have positive spatial components such that the raising and lowering of
indices only changes the sign of the temporal components. The components of the field strength

14
tensor {F }
µν
are related to the components of the spacetime potential
{ } { }
Aµ = A0 ≡ V / c, A1 , A2 , A3 via

F µν = ∂ µ Aν − ∂ν Aµ . (37)

The field strength tensor F is naturally divided into temporal and spatial components. The three
independent temporal components are associated with the vector electric field

E ' / c ≡ F 0i = ∂ 0 Ai − ∂ i A0 (38)

while the three spatial components naturally form a 3× 3 second rank anti-symmetric tensor
magnetic field

B ij ≡ F ij = ∂ i A j − ∂ j Ai . (39)

Maxwell’s inhomogeneous equations, in vacuum, are expressed concisely in terms of the field
strength tensor:

∂ν F µν = µ0 J µ (40)

where the spacetime current density has components J µ = J 0 ≡ cρ , J 1 , J 2 , J 3 .

r
{ }
In (3+1)-D, ρ is the volume charge density and J = J1 xˆ1 + J 2 xˆ2 + J 3 xˆ3 represents the distribution
of the current over a cross-sectional area. Maxwell’s homogeneous equations follow from the
relationship between the fields and the spacetime potential (eqs. 38-39). In terms of the vector
electric field E and tensor magnetic field B, Maxwell’s equations in differential form are:

ρ
∂ i Ei = , ∂ i E j − ∂ j Ei = −c∂ 0 Bij ,
ε0
∂0E j
ε ijk ∂ i B jk = 0 , − ∂ i Bij = + µ0 J j . (41)
c

We do not distinguish between covariant and controvariant spatial indices {i, j , k }∈ {1,2,3} since
only raising or lowering of temporal components involves a sign change. With the line element ds
r
expressed as an anti-symmetric second-rank tensor and the differential area dA expressed as a
vector via the following equation

1
Cij = ε ijk Ck , Ck = ε ijk Cij ,
2!

the integral form of Maxwell’s equations are:

1
∫ E dA = ε ∫ ρdV ,
S
i i
0 Venc
ε ijk ∫ Ei ds jk = −cε ijk ∂ 0 ∫ Bij dAk ,
C S enc

 ∂ 0 Ei 
ε ijk ∫ Bij dAk = 0 , ∫ B ds
ij ij = 2! ∫  µ J 0 i + dAi . (42)
S C S enc
c 

15
When generalizing eq. (42) to (N+1)-D, the higher-dimensional analog of the differential line
element ds will have N − 2 spatial dimensions and the higher-dimensional analog of the
r
differential area dA will have N − 1 spatial dimensions. It is also useful to work with an anti-
symmetric second-rank tensor G that is dual to F:

1 µνρσ
G µν = ε Fρσ . (43)
2!

Maxwell’s homogeneous equations are obtained from the dual tensor G via

∂ν G µν = 0 . (44)

Alternatively, Maxwell’s homogeneous equations may be extracted from

ε λµνρ Tλµν = 0 , (45)

where the anti-symmetric third-rank tensor T is defined as

Tλµν ≡ ∂ λ Fµν . (46)

The most straightforward generalization of Maxwell’s equations to (4+1)-D begins by extending

eq. (38) to

∂ Γ F ΦΓ = µ 4 +1 J Φ (47)

where the uppercase Greek letters {Φ, Γ}∈ {0,1,2,3,4} represent (4+1)-D spacetime indices and the
(4+1)-D permittivity and permeability of free space ε 4+1 and µ4+1 , respectively, have different
dimensions than their (3+1)-D counterparts ε 0 and µ0 .
The 5 × 5 (4+1)-D field strength tensor F has 10 independent components, which naturally divides
into a 4-component vector electric field

E I / c = F 0 I = ∂ 0 A I − ∂ I A0 (48)

and a 4 × 4 anti-symmetric second rank tensor magnetic field

B IJ = F IJ = ∂ I A J − ∂ J A I (49)
r
with 6 independent components. In terms of the vector electric field E and tensor magnetic field
B, Maxwell’s equations in (4+1)-D in differential form become

ρ
∂ I EI = , ∂ I E J − ∂ J EI = −c∂ 0 BIJ ,
ε 4 +1
∂ 0 EJ
ε IJKL ∂ I BJK = 0 , − ∂ I BIJ = + µ 4 +1 J J . (50)
c

16
Gauss’s law in magnetism, which states that the net magnetic flux is always zero as a
consequence that magnetic monopoles have not been observed, can alternatively be expressed in
differential form as:

∂ I BJK + ∂ J BKI + ∂ K BIJ = 0 . (51)

r
Alternatively, the differential volume may also be expressed as a vector dV via the following
equation

1
CIJK = ε IJKLCL , CL = ε IJKLCIJK .
3!

With the fields and differential elements expressed in terms of appropriate vectors and tensors,
(4+1)-D Maxwell’s equations can be expressed in integral form as:

1
∫ E dV ρdW , ε IJKL ∫ EI dAJK = −cε IJKL ∂ 0 ∫ BIJ dVK ,
ε 4 +1 W∫
I I =
V enc S Venc

 ∂ E 
ε IJKL ∫ BIJ dVK = 0 , ∫B IJ dAIJ = 2! ∫  µ 4 +1 J I + 0 I dVI . (52)
Venc 
V S
c 

While electric flux remains a scalar, magnetic flux has four components in (4+1)-D:

Φ e = ∫ EI dVI , Φ mL = ε IJKL ∫ BIJ dVK . (53)

V V

In (4+1)-D, the continuity equation reads

∂ I J I = −c∂ 0 ρ . (54)

In (4+1)-D, the dual tensor G is an anti-symmetric third-rank tensor:

1 ΦΓΛΣΩ
G ΦΓΛ ≡ ε FΣΩ . (55)
3!

Maxwell’s homogeneous equations may be expressed in terms of the dual tensor G via

∂ ΛG ΦΓΛ = 0 (56)

or, equivalently, in terms of an anti-symmetric third-rank tensor T via

ε ΦΓΛΣΩTΛΣΩ = 0 (57)

where T is defined as

TΦΓΛ ≡ ∂ Φ FΓΛ . (58)

17
In (5+1)-D, the differential form of Maxwell’s inhomogeneous equations are expressed concisely in
terms of the 6 × 6 field strength tensor F as

∂ Γ5+1 F Φ 5+1Γ5+1 = µ5 +1 J Φ 5+1 (59)

where {Φ 5 +1 , Γ5 +1}∈ {0,1,2,3,4,5} represent (5+1)-D spacetime indices. Maxwell’s homogeneous

equations can be expressed in terms of the dual tensor G or, equivalently, its counterpart T, as

∂ Ξ 5+1 G Φ 5+1Γ5+1Λ 5+1Ξ 5+1 = 0 , εΦ 5+1Γ5+1 Λ 5+1Σ 5+1Ω 5+1Ξ 5+1

TΛ 5+1Σ 5+1Ω 5+1 = 0 , (60)

where

1 Φ 5+1Γ5+1Λ 5+1Σ 5+1Ω 5+1Ξ 5+1

G Φ 5+1Γ5+1Λ 5+1Σ 5+1 ≡ ε FΩ 5+1Ξ 5+1 , TΦ 5+1Γ5+1Λ 5+1 ≡ ∂ Φ 5+1 FΓ5+1 Λ 5+1 . (61)
4!

In terms of the 5-component vector electric field E and the 5 × 5 tensor magnetic field B, the
differential form of Maxwell’s equations in (5+1)-D are:

ρ
∂ I 5 EI 5 = , ∂ I 5 E J 5 − ∂ J 5 EI 5 = −c∂ 0 BI 5 J 5 ,
ε 5 +1
∂ 0 EJ 5
εI 5 J 5 K 5 L5 M 5
∂ I 5 BJ 5 K 5 = 0 , − ∂ I 5 BI 5 J 5 = + µ5 +1 J J 5 , (62)
c

where {I 5 , J 5 } = {1,2,3,4,5} . In integral form, Maxwell’s equations in (5+1)-D are:

1
∫E ρdX , εI dVJ 5 K 5 = −cε I 5 J 5 K 5 L5 M 5 ∂ 0
ε 5 +1 X∫ ∫E ∫B
I5 dWI 5 = 5 J 5 K 5 L5 M 5 I5 I5 J5 dWK 5 ,
W enc V Wenc

 ∂ E 
εI 5 J 5 K 5 L5 M 5 ∫B I5 J5 dWK 5 = 0 , ∫B I5J5 dVI 5 J 5 = 2! ∫  µ5 +1 J I 5 + 0 I 5 dWI 5 , (63)
W V Wenc 
c 

where dX is the differential five-dimensional volume element.

For an isolated point charge in (4+1)-D, the electric field lines radiate outward in four-dimensional
r
space. For a Gaussian hypersphere of radius r , the electric field E is parallel to the differential
r
volume element dV everywhere on the three-dimensional volume x12 + x22 + x32 + x42 = r 2
bounding the (four-dimensional) hypersphere defined by x12 + x22 + x32 + x42 ≤ r 2 . In (4+1)-D, the
r
electric field E due to a point charge is thus found to be

1 q r q
∫ EI dVI = E 2π r = ∫ ρdW = E=
2 3
, rˆ , (64)
V
ε 4 +1 W enc
ε 4 +1 2π ε 4 +1r 3
2

thence, we obtain

18
 r
1 E 2π 2 r 3
∫V EI dVI = E 2π r = ε 4 +1 W∫ ρdW = rˆ . (64b)
2 3

enc

N
We have utilized the result that the (N–1)-dimensional volume ∑x
i =1
2
i = r 2 bounding a solid sphere
N
in N-dimensional space ∑x
i =1
2
i ≤ r 2 is

2π N / 2 r N −1
VNsphere = (65)
Γ( N / 2 )
−1

where Γ( z ) is the gamma function.

r
Generalizing to (N+1)-D, the electric field E due to a point charge is

r Γ(N / 2) q
E= rˆ (66)
2π N /2
ε N +1r N −1

where the (N+1)-D permittivity of free space ε N +1 has different dimensionality than its (3+1)-D
counterpart ε 0 :

[ε N +1 ] = [εN0−]3
[L ] . (67)

Thus, Coulomb’s law for the force exerted by one point charge q1 on another point charge q2
r
displaced by the relative position vector R21 away from the first is a 1 / r N −1 force law:

r Γ( N / 2 ) q1q2 ˆ
Fq 2 , q1 = R21 . (68)
2π N / 2 ε N +1R21N −1

r r
In (N+1)-D, the electric field E (r ) at a field point located at r may alternatively be derived
r
through direct integration over the differential source element dq ' :

Γ(N / 2)
q'
r r
E (r ) =
1 ˆ
∫
2π ε N +1 0 R N −1
N /2
Rdq' (69)

r r r
where the relative position vector R = r − r ' extends from the differential source element dq '
r r
located at r ' to the field point located at r .
The (N+1)-D electric field is related to the (N+1)-D scalar potential via
r
r r ∂A
E = −∇V − . (70)
∂t

In electrostatics, the (N+1)-D scalar potential is derived from the following integral:

19
 Γ( N / 2 )
q'

V (r ) =
1
∫
2π ( N − 2)ε N +1 0 R
N/2 N −2
dq' . (71)

The traditional Ampèrian loop is promoted to a two-dimensional Ampèrian surface in (4+1)-D. A

suitable choice for a steady filamentary current I 0 running along the x1 -axis in the two-
dimensional surface x22 + x32 + x42 = r 2 of a solid sphere x22 + x32 + x42 ≤ r 2 in the three-dimensional
subspace x2 x3 x4 . In this case, the only non-vanishing components of the magnetic field tensor B
are {B1I } and {BI 1} . By symmetry, the magnetic field scalar, defined by the tensor contraction
B ≡ BIJ BIJ / 2! , is constant everywhere on the surface x22 + x32 + x42 = r 2 and the components of
the magnetic field tensor are related by B1I = B1J = − BI 1 = − BJ 1 , where I ≠ J , such that
B = B122 + B132 + B142 = B12 3 . However, it is simpler to work in spherical coordinates, where
BIJ dAIJ = 2 Br 2 dΩ . Thus, application of Ampère’s law yields

µ 4 +1I 0
∫B dAIJ = 2 ∫ Br 2 dΩ = B8πr 2 = 2 µ 4 +1I enc , B= . (72)
4πr 2
IJ
S S

In (3+1)-D, the magnetic flux Φ m is a scalar and the magnetic field lines for this infinite steady
filamentary current I 0 are traditionally drawn as a concentric circles x22 + x32 = r 2 in the x2 x3
plane. However, in (4+1)-D the magnetic flux is a four-component vector Φ mI according to eq. { }
(53). This corresponds to the fact that circles x22 + x32 = r 2 , x22 + x42 = r 2 , and x32 + x42 = r 2 lying in
the x2 x3 , x2 x4 , and x3 x4 planes, respectively, are all orthogonal to the current running along the
x1 -axis. In (N+1)-D, the magnetic flux becomes an anti-symmetric (N-3)-rank tensor.
Generalizing to (N+1)-D, the magnetic field due to a steady filamentary current I 0 is

 N − 1  µ 4 +1I 0
B (r ) =
1
( N −1) / 2 Γ  N − 2 . (73)
2π  2  r

In (N+1)-D, the force per unit length l that one steady filamentary current I1 exerts on a parallel
r
steady filamentary current I 2 separated by the relative position vector R21 is:
r
FI 2 , I1 1  N − 1  µ 4 +1I1I 2 ˆ
= ( N −1) / 2 Γ  N − 2 R21 . (74)
l 2π  2  R21

In (4+1)-D, the tensor magnetic field B(r ) at a field point located at r may alternatively be
r r

derived through direct integration. For a steady filamentary current,

 N −1
BIJ (r ) =
r 1 1
( N −1) / 2 Γ  µ N +1I 0 ∫ 4 RJ dsI − RI ds J (75)
4π  2  R

20
 r r r r
where the relative position vector R = r − r ' extends from the differential source element ds '
r r
located at r ' to the field point located at r . If the current is instead distributed over a one-,two,
or three-dimensional cross section, the tensor magnetic field B(r ) is
r

 N −1  N −1
BIJ (r ) =  µ N +1 ∫ 3 J1dAIJ ds , BIJ (r ) = ( N −1) / 2 Γ
r 1 1 r 1 1
Γ
( N −1) / 2  µ N +1 ∫ 3 J 2 dAIJ ds ,
4π  2  R 4π  2  R
 N −1
BIJ (r ) = ( N −1) / 2 Γ  µ N +1 ∫ 4 J 3 (RJ dVI − RI dVJ )ds , (76)
r 1 1
4π  2  R

where J1 , J 2 , and J 3 are the corresponding to the current densities.

With a single non-compact extra dimension x4 , two (4+1)-D electric charges q1 and q2
communicate via the exchange of a (4+1)-D photon. Classically, there is a single straight-line path
with which the (4+1)-D photon can reach q2 from another q1 . One result of this single non-
compact extra dimension x4 is that Coulomb’s law becomes a 1 / r 3 force law. If instead the extra
dimension x4 is compactified on a torus with radius R , there exists an infinite discrete set of
paths with which a classical (4+1)-D photon could travel from q1 to q2 .
The net force exerted on q2 is the superposition of the (4+1)-D 1 / r 3 Coulomb forces (eq. 68)
associated with each path:

r ∞
1 1 ˆ
F=
2π ε 4 +1
2
q1q2 ∑R
n = −∞
3
Rn (77)
n

where Rn = R32 + (∆x4 + 4πnR ) is the path length corresponding to a classical (4+1)-D photon
2

R Rˆ + ∆x4 xˆ4
winding around the torus n times and Rˆ n = 3 3 is a unit vector directed from q1 to q2
Rn
(and R̂3 is a unit vector along the axis of the torus). In terms of its three-dimensional and extra-
dimensional components, the net force is

r 1 ∞
R3 Rˆ3 + ∆x4 xˆ4
F= ∑
[R ]
q1q2 . (78)
2π 2ε 4 +1 + (∆x4 + 4πnR )
2 2 2
n = −∞
3

The net force can also be expressed in terms of the (4+1)-D scalar potential V (R4 ) via
r r
F = − q2∇V . (79)

The (4+1)-D scalar potential due to the source q1 at the location of q2 is

∞
V (R4 ) =
1 1
q1 ∑ . (80)
4π ε 4 +1 n = −∞ R + (∆x4 + 4πnR )
2 2 2
3

Note that the (4+1)-D scalar potential is periodic in the extra dimension:

21
 V (R3 , ∆x4 ) = V (R3 , ∆x4 ± 2πnR ) . (81)

In the limit that the charges are very close compared to the radius of the extra dimension, i.e.
R3 << R and ∆x4 << R , the n = 0 term dominates and the net force is approximately the (4+1)-D
1 / r 3 Coulomb force:

r 1
Fq1 , q 2 ≈ q1q2 Rˆ 4 . (82)
2π ε 4 +1R43
2

This is the limit where the underlying (4+1)-D theory of electrodynamics – i.e. Maxwell’s equations
with a non-compact extra dimension – governs the motion. In the opposite extreme, where the
compact extra dimension is very small compared to the separation of the charges, i.e. R3 >> R ,
an integral provides a good approximation for the sum:

∞
V (R4 ) ≈
1 1 1
8π ε 4 +1R
3
q1 ∫
y = −∞
R +y 22
3
dy = 2
8π ε 4 +1RR3
q1 . (83)

This is the limit where the effective (3+1)-D theory of electrodynamics – i.e. the usual (3+1)-D form
of Maxwell’s equations – governs the motion. In order for the effective (3+1)-D scalar potential in
eq. (83) to be consistent with the usual (3+1)-D form of Coulomb’s law, it is necessary that the
(4+1)-D permittivity of free space ε 4+1 be related to the usual (3+1)-D permittivity ε 0 via

ε0
ε 4 +1 = . (84)
2πR

It follows that

µ 4 +1 = 2πµ0 R (85)

such that an electromagnetic wave propagates at the usual speed of light in vacuum

1 1
c= = . (86)
ε 4 +1 µ 4 +1 ε 0 µ0

In (N+1)-D, Coulomb’s law is a 1 / r N −1 force law if the extra dimensions are non-compact. If
instead there is toroidal compactification and the extra dimensions are symmetric . i.e. they all
have the same radius R - then, in the limit that R3 >> R , Coulomb’s law is approximately a 1 / r 2
force law in the effective (3+1)-D theory. In this case, the (N+1)-D permittivity of free space ε N +1 is
related to the usual (3+1)-D permittivity ε 0 via

ε0
ε N +1 = (87)
(2R π ) N −3  N −3
Γ 
 2 

22
where N > 3 .
Thence, the eq. (71), for the eq. (87), can be written also as follow:

Γ( N / 2)
(2R π ) N −3 N − 3
Γ  q'
V (r ) =  2  1
2π N / 2 ( N − 2 ) ε0 ∫
0
R N −2
dq ' , (87b)

where Γ( z ) is the gamma function.

The deviation in the usual 1 / r 2 form of Coulomb’s law can be computed for a specified geometry
of extra dimensions through the lowest-order corrections to the integration in eq. (83). The
effects are similar to the deviation in the usual 1 / r 2 form of Newton’s law of universal gravitation.

3. On some equations concerning the noncommutativity in String Theory: the Dirac-Born-

Infeld action, connections with the Maxwell electrodynamics and the Maxwell's equations,
noncommutative open string actions and D-brane actions [3] [4] [5] [6] [7]

3.1 The Dirac-Born-Infeld action: connections with the Maxwell’s equations

To obtain the low-energy effective action for the gauge field, we need to expand the worldsheet
theory about a background field X i , and compute the (divergent) one-loop counterterm:

− i ∫ dτΓi ( A( X ), Λ )∂τ X i (88)

where Λ is an ultraviolet cutoff. Setting Γi to zero gives a condition on the gauge field Ai ( X ) ,
equivalent to worldsheet conformal invariance, or vanishing of the β -function.
That gives the spacetime equation of motion, from which the action can be reconstructed.
Performing the background field expansion

X i = X i + ξ i (89)

where X i is an arbitrary classical solution of the worldsheet equations of motion, we find:

δS δ 2S
S ( X ) = S (X ) + ∫ i
1
ξ + ∫ i
i
ξ iξ j + ... (90)
δX X =X 2 δX δX j
X =X

The linear term in ξ i vanishes as X is a solution of the equation of motion. The quadratic term is
easily evaluated:

1 δ 2S
2 ∫ δX iδX j
X =X
ξ iξ j =
1
[ dσdτg ∂ ξ ∂ ξ
4πα ' ∫
ij a
i
a
j
( )]
+ i ∫ dτ ∂ i Fjk ∂τ X jξ iξ k + Fijξ j ∂τ ξ i . (91)

The correction to the worldsheet action is:

23
 K ik (τ ,τ ') τ =τ ' , (92)
i
dτ∂ F ∂τ X
4πα ' ∫
j
i jk

where we have ignored possible UV finite terms.

Recalling the formula for the boundary propagator, we have:

lim K ij (τ ;τ ') = −2α ' G ij (F )ln Λ + (finite) (93)

τ →τ '

and hence the equation of motion is:

ik
 1 1 
∂ i Fjk G (F ) ≡ ∂ i Fjk 
ik
g  = 0 . (94)
g +F g−F

When we think of G ij (F ) as the (inverse) open-string metric, then this looks just like the free
Maxwell’ equations. It turns out that the desired open-string effective action is:

[ ]
S NS − NS Ai ; g ij , Bij =
1
gs ∫
d 10 x det ( g + F ) . (95)

It is possible to show that:

 δ 
∂i  det ( g + F )  = − det ( g + F )G jk (F )∂ i Fkl G li (F ) . (96)
 δ (∂ i A j ) 

Since the factor det ( g + F ) is nonzero and the matrix G jk (F ) is invertible, it follows that setting
the above expression to zero is equivalent to:

∂ i Fkl G li (F ) = 0 (97)

which is the desired equation of motion. At lowest order in F this is equivalent to the Maxwell
equations:

∂ i Fkl g li = 0 (98)

but in general, as we noted, it has nonlinear corrections.

The action:

[ ]
S NS − NS Ai ; g ij , Bij =
1
gs ∫ d 10 x det ( g + F ) = ∫ d 10 x det ( g + 2πα ' (B + F ))
1
gs
(99)

is called the Dirac-Born-Infeld (DBI) action. Expanding this action to quadratic order in F , it is
easily seen that it is proportional to the usual action of free Maxwell electrodynamics:

24
 [ ]
S NS − NS Ai ; g ij , Bij ≈ ∫ Fij F ij + ... (100)

i.e. we can write the eq. (100) also as follow:

[ ]
S NS − NS Ai ; g ij , Bij ≈ ∫ Fij F ij + ... ≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0
C
d
dt ∫S
E ⋅ nˆdA , (100b)

thence, we obtain the following interesting mathematical connection:

[ ]
S NS − NS Ai ; g ij , Bij =
1
gs ∫ d 10 x det ( g + F ) = ∫ d 10 x det ( g + 2πα ' (B + F )) ≈
1
gs
d
≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA . (100c)
C dt S

This is consistent with the fact that the linearized equations of motion are just Maxwell’s
equations.
We will now recast the DBI action in a different form. For this, let us first define

ij
 
G ≡ G (F = 0 ) = 
1 1
ij ij
g  (101)
 g + 2πα ' B g − 2πα ' B 
θ ij (F = 0) 
ij
θ ij 1 1 
≡ = − 2πα ' B  . (102)
2πα ' 2πα '  g + 2πα ' B g − 2πα ' B 

We abbreviate G ij by G −1 . We also define the matrix Gij , abbreviated G , to be the matrix inverse
of G ij . Thus we have defined two new constant tensors G −1 ,θ in terms of the original constant
tensors g, B . In particular,

ij
θ  
ij
 −1 1 
G +  =   . (103)
 2πα '   g + 2πα ' B 

It is illuminating to rewrite the DBI Lagrangian in terms of the new tensors. This is achieved by
writing:

 
 
det ( g + 2πα ' (B + F )) =
1 1 1
det  + 2πα ' F  =
gs gs  G −1 + θ 
 
 2πα ' 
det (1 + θF )  1 
det (G (1 + θF ) + 2πα ' F ) =
1 1 1
= det  G + 2πα ' F .
gs  Gθ  gs  Gθ   1 + θF 
det 1 +  det 1 + 
 2πα '   2πα ' 
(104)
Defining

25
 1  Gθ 
Fˆ = F , Gs = g s det 1 +  (105)
1 + θF  2πα ' 

we end up with the relation

1
gs
det ( g + 2πα ' (B + F )) =
1
Gs
det (1 + θF ) det G + 2πα ' Fˆ . (106)( )

The relation between F and F̂ can be easily inverted, leading to:

1
F = Fˆ (107)
1 − θFˆ

from which it also follows that:

1
1 + θF = . (108)
1 − θFˆ

Hence we have:

1
det ( g + 2πα ' (B + F )) =
1 1
( )
det G + 2πα ' Fˆ . (109)
gs Gs (
det 1 − θFˆ )
In what follows, we must be careful to remember that the above equations were obtained in the
strict DBI approximation of constant F .
Apart from the factor ( )
det 1 − θF̂ in the denominator, the right hand-side looks like a DBI
Lagrangian with a new string coupling Gs , metric Gij and gauge field strength F̂ , and no B -
field. Let us therefore tentatively define the action:

SˆDBI =
1
Gs ∫ ( )
det G + 2πα ' Fˆ . (110)

Let us start by expanding the relation through which we defined F̂ , to lowest order in θ :

( )
Fˆij = Fij − Fikθ kl Flj + Ο θ 2 . (111)

Inserting the definition of Fij , we get:

( )
Fˆij = ∂ i A j − ∂ j Ai + θ kl (∂ i Ak ∂ j Al − ∂ i Ak ∂ l A j − ∂ k Ai ∂ j Al + ∂ k Ai ∂ l A j ) + Ο θ 2 . (112)

We can make a nonlinear redefinition of Ai to this order, which absorbs three of the four terms
linear in θ :

26
  1 
( )
Aˆi = Ai − θ kl  Ak ∂ l Ai + Ak ∂ i Al  + Ο θ 2 . (113)
 2 

We can find that

Fˆij = ∂ i Aˆ j − ∂ j Aˆi + θ kl ∂ k Aˆi ∂ l Aˆ j . (114)

It is clear that there is no further redefinition of  that will absorb the last term. However, we
note that this term is:
{ }
θ kl ∂ k Aˆi ∂ l Aˆ j = Aˆi , Aˆ j (115)

where {...,} is the Poisson bracket with Poisson structure θ . Thus, to linear order in θ , we have
found that:
{ }
Fˆij = ∂ i Aˆ j − ∂ j Aˆi + Aˆi , Aˆ j . (116)

This looks like a non-Abelian gauge field strength, except that there is a Poisson bracket instead
of a commutator.
We can say that the field strength F̂ij is a noncommutative field strength related to its gauge
potential Âi by:
[ ]
Fˆij = ∂ i Aˆ j − ∂ j Aˆi − i Aˆi , Aˆ j ∗ . (117)

The map
k
 1 
Fˆij = Fik   j (118)
 1 + θF 
 1 
Aˆi = Ai − θ kl  Ak ∂ l Ai + Ak ∂ i Al  + Ο θ 2 ( ) (119)
 2 

is known as the Seiberg-Witten map.

We remember that the Seiberg-Witten gauge theory is a set of calculations that determine the
low-energy physics — namely the moduli space and the masses of electrically and magnetically
charged supersymmetric particles as a function of the moduli space.

This is possible and nontrivial in gauge theory with N = 2 extended supersymmetry by combining
the fact that various parameters of the Lagrangian are holomorphic functions (a consequence of
supersymmetry) and the known behavior of the theory in the classical limit.

The moduli space in the full quantum theory has a slightly different structure from that in the
classical theory

The noncommutative description is parametrized by the noncommutativity parameter θ , the

open-string metric Gij , the open-string coupling Gs , and a “description parameter” Φ , in terms
of which the relationship between closed-string and open-string parameters is given by:

27
 det ( g + 2πα ' B ) det (G + 2πα ' Φ )
ij
 1  θ 1
N ≡ 
ij
 = + , = . (120)
 g + 2πα ' B  2πα ' G + 2πα ' Φ gs Gs

Now we wish to compare the sum of the commutative DBI action S DBI plus the derivative
corrections to it ∆S DBI with the noncommutative DBI action Ŝ DBI , after taking the Seiberg-Witten
limit on both sides.
The dilaton couples to the entire Lagrangian density, so we need to consider the full DBI action.
We will start by restricting to terms quadratic in F . To this order, we have:

det ( g + 2πα ' B )  2πα ' (2πα ')

2
(2πα ')
2

S DBI = ∫ 1 + tr ( NF ) − tr ( NFNF ) + (trNF )2 + ... . (121)
gs  2 4 8 

θ ij
In the Seiberg-Witten limit we have N ij → and therefore:
2πα '

det ( g + 2πα ' B )  1 

1 + tr (θF ) − tr (θFθF ) + (trθF ) + ... . (122)
1 1
=∫
2
S DBI SW 
gs  2 4 8 

Let us now convert the commutative field strengths F appearing in this expression into
noncommutative field strengths F̂ , using the Seiberg-Witten map. To the order that we need it,
this map is:
F = Fˆ + θ kl Aˆ , ∂ Fˆ
ab − Fˆ , Fˆ
ab ((123)k l ab
∗2
ak bl
∗2
)
where
Fˆab = ∂ a Aˆb − ∂ b Aˆ a + θ kl ∂ k Aˆ a , ∂ l Aˆb . (124)
∗2

Here we have used an identity relating the Moyal ∗ commutator and the ∗2 product:

− i[ f , g ]∗ = θ ij ∂ i f , ∂ i g ∗2
. (125)

Inserting the Seiberg-Witten map into eq. (122), we find

det ( g + 2πα ' B )  1

S DBI SW =∫ 1 + θ ij ∂ j Aˆi + θ baθ kl ∂ k Aˆ a , ∂ l Aˆb +
gs  2 ∗2

1
(
+ θ abθ kl Aˆ k , ∂ l Fˆab
2 ∗2
− Fˆak , Fˆbl
∗2
)− 14 θ θij kl
(
1
8
)
Fˆ jk Fˆli + θ ij Fˆij  . (126)
2



Some manipulation of the last few terms permits us to rewrite this as:

det ( g + 2πα ' B )  1 1

S DBI SW =∫ 1 + θ ij ∂ j Aˆi + θ ijθ kl ∂ j Aˆ k , ∂ l Aˆi + θ baθ kl Aˆ k , ∂ l Fˆab +
gs  2 ∗2 2 ∗2

28
 1
4
(
+ θ ijθ kl Fˆ jk , Fˆli
∗2
)
1
− Fˆ jk Fˆli + θ ijθ kl Fˆ ji , Fˆlk
8 ∗2
1
(
− θ ijθ kl Fˆ ji , Fˆlk
8 ∗2

)
− Fˆ ji Fˆlk  (127)


which is the form in which it will be useful.

Let us now turn to the noncommutative side. Here, we only need to keep the terms arising from
expansion of the Wilson line, since all other terms are suppressed by powers of α ' in the Seiberg-
Witten limit. The Wilson line give us:

det (G + 2πα ' Φ )  1 

SˆDBI SW =∫ 1 + θ ij ∂ j Aˆi + θ ijθ kl ∂ j ∂ l Aˆi , Aˆ k  . (128)
Gs  2 ∗2

After some rearrangements of terms, this can be written:

det (G + 2πα ' Φ )  1

SˆDBI SW =∫ 1 + θ ij ∂ j Aˆi + θ ijθ kl ∂ j Aˆ k , ∂ l Aˆi +
Gs  2 ∗2

1 1 
+ θ baθ kl Aˆ k , ∂ l Fˆab + θ ijθ kl Fˆ ji , Fˆlk  . (129)
2 ∗2 8 ∗2

Now we can take the difference of eqs. (129) and (127). The prefactor in front of each expression is
the same, by virtue of eq. (120). Apart from this factor and the integral sign, the result is:

SˆDBI SW − S DBI SW
1
4
(
= θ ijθ kl Fˆ jk , Fˆli
∗2
) 1
8
(
− Fˆ jk Fˆli − θ ijθ kl Fˆ ji , Fˆlk
∗2
)
− Fˆ ji Fˆlk . (130)

To the order in which we are working, we may replace F̂ by F everywhere in this expression.
This, then, is our prediction for the correction ∆S DBI , to order (α ') and to quadratic order in the
2

field strength F , after taking the Seiberg-Witten limit. We note that this is manifestly a higher-
derivative correction: it vanishes for constant F , for which the ∗2 product reduces to the ordinary
product. Expanding the ∗2 product to 4-derivative order, we find that

1  ij kl mn rs 1 
∆S DBI SW =− θ θ θ θ ∂ m ∂ r F jk ∂ n ∂ s Fli − θ ijθ klθ mnθ rs ∂ m∂ r Fji ∂ n ∂ s Flk  (131)
96  2 

which gives:

∆S DBI
(2πα ')4 hij h kl h mn h rs ∂ 1 
SW = −  m ∂ r F jk ∂ n∂ s Fli − h ij h kl h mn h rs ∂ m∂ r F ji ∂ n ∂ s Flk  (132)
96 2 

where the matrix h ij is defined as:

ij
 1 
h ≡ 
ij
 . (133)
 g + 2πα ' (B + F ) 

29
Taking the Seiberg-Witten limit, which amounts to the replacement 2πα ' h → (1 + θF ) θ , and
−1

further restricting to terms quadratic in F , we find exact agreement with eq. (131) above.
Now we will compare the coupling of the bulk graviton to the energy-momentum tensor on the
commutative and noncommutative sides. On the commutative side, we start again with the
expression in eq. (121), but this time we use the full form of N as defined in eq. (120):

ij
 1  θ ij
N ≡ 
ij
 = + M ij (134)
 g + 2πα ' (B + F )  2πα '

where
ij
 1 
M ≡
ij
 . (135)
 G + 2πα ' Φ 

Thence, we obtain:

ij
θ ij
ij
 1   1 
N ≡ 
ij
 = +  . (135b)
 g + 2πα ' (B + F )  2πα '  G + 2πα ' Φ 

As the linear coupling to the graviton starts at order (α ') , we now have to go beyond the leading
2

term in the Seiberg-Witten limit. Hence we will keep terms up to order M 2 . Expanding S DBI
around this limit and keeping terms to order (α ') , and using the Seiberg-Witten map, we find:
2

det ( g + 2πα ' B )  2πα '

{ ( )}+ (2πα8 ')
2
S DBI = ∫ 1 + M ji Fij + θ kl Ak , ∂ l Fij + θ kl F jk , Fli ⋅
gs 2 ∗2 ∗2


(trMF )(trθF ) + (trMF )2  − (2πα ') trMFMF + 2 trMFθF  + terms not involving M +
2
 2
⋅
 2πα '  4  2πα ' 
]3
order F . (136)

Turning now to the noncommutative action, the graviton coupling is obtained by expanding the
DBI action around the Seiberg-Witten limit to order (α ') . We have, in momentum space:
2

SˆDBI =
1
Gs ∫ 
( ( ))
L∗  det G + 2πα ' Fˆ + Φ W ( x, C ) ∗ eik . x =


=
1
∫
 1 2
( ) (
 
)
det G L∗ 1 − (2πα ') trG −1 Fˆ + Φ G −1 Fˆ + Φ W ( x, C ) ∗ eik . x + ... (137)
Gs  4  

The piece of the above expression that is order 1 in α ' has already been computed earlier for the
dilaton coupling. It contributes to the coupling of the trace of the graviton. The new non trivial
coupling is given by the order (α ') term.
2

To compare with the commutative side, it is convenient to expand the above action differently, in
terms of M rather than G . We get:

30
  2πα '  
det (G + 2πα ' Φ )  L∗ (W ( x, C )) +
1 1
SˆDBI = ∫ trMF + M θ ∂ j Flk , Ai ∗2 + trMF , trθF +
kl ij
∗2
Gs  2  2 

−
(2πα ')2 tr MF , MF + (2πα ')2 trMF , trMF + ... . (138)
∗2 ∗2 
4 8 

Now taking the difference of the noncommutative and commutative actions in eqs. (138) and
(136), and expanding the result to 4-derivative order, we get the prediction:

2πα '  ij kl mn rs 1 ij kl mn rs 
∆S DBI SW =− M θ θ θ ∂ m ∂ r F jk ∂ n ∂ s Fli − M θ θ θ ∂ m ∂ r F ji ∂ n ∂ s Flk  +
48  2 

−
(2πα ')2 M ij M klθ mnθ rs ∂ 1 
∂ r F jk ∂ n∂ s Fli − M ij M klθ mnθ rs ∂ m ∂ r F ji ∂ n∂ s Flk  . (139)
 m
96  2 

Note that contrary to appearances, both of the above terms are of order (α ') . This is because if
2

one inserts G in place of M in the first line, the result vanishes.

3.2 The noncommutative open string actions, Chern-Simons theory and D-brane actions

Let us focus on a particular Chern-Simons coupling, the one involving the Ramond-Ramond 6-
form C (6 ) . In the commutative theory, this is just

1 (6 )
2∫
C ∧ F ∧ F . (140)

The noncommutative version of this coupling is obtained from the commutative one by making
the replacement:
1
F → Fˆ (141)
1 − θFˆ

where F̂ is the noncommutative gauge field strength, multiplying the action by a factor
( )
det 1 − θF̂ , and using the Moyal ∗ -product defined in the following equation:

is r
∂ pθ ∂q
f ( x ) ∗ g ( x ) ≡ f ( x )e 2 g ( x ) . (142)
pq

To make a coupling that is gauge-invariant even for nonconstant fields, this has to be combined
with an open Wilson line. The resulting expression for the coupling to C (6 ) is more conveniently
expressed in momentum space, where C (6 ) (k ) is the Fourier transform of C (6 ) ( x ) :
~


(

C (− k ) ∧ ∫ L∗  det 1 − θFˆ  Fˆ
1 ~ (6 )
)
1  ˆ 1 
 ∧F
 ik . x
W (x, C ) ∗ e . (143)
2   1 − θF   1 − θF 
ˆ ˆ 

31
Here W ( x, C ) is an open Wilson line, and L∗ is the prescription of smearing local operators along
the Wilson line and path-ordering with respect to the Moyal product. Evaluation of the L∗
prescription leads to ∗n products.
The above expression can easily be re-expressed in position space, and it turns into:

C ∧ (F ∧ F )∗2 . (144)
1 (6 )
2∫

Let us denote the sum of all derivative corrections to SCS as ∆SCS . Our starting point is the
expression
dσdθDφ i Ai (φ )
i
− ∫
SCS + ∆SCS = C e 2πα '
B (145)
R

where C represents the RR field, and B R

is the Ramond-sector boundary state for zero field
strength. We are using superspace notation, for example φ i = X i + θψ i and D is the
supercovariant derivative. It is possible write the following expression:

∫ [ ]∑
∞ k +1
∑k =0 (k +1)! k + 2 Dφ jφ iφ a1 ...φ ak ∂ a1 ...∂ ak Fij ( x )
i 1 ~ ~ ~ ~ ∞
∫ X .. X ∂ a1 ..∂ ak Fij ( x )
dσdθ i ~ 1 ~ a1 ~ ak
dσ Ψ iψ 0j +ψ 0iψ 0j
2πα '
SCS + ∆SCS = C e ×e 2πα ' k =0 k!
B
R
(146)
where nonzero modes have a tilde on them, while the zero modes are explicitly indicated.
We can drop the first exponential factor in eq. (146) above, as well as the first fermion bilinear
~
Ψ iψ 0j in the second exponential. Then, expanding the exponential to second order, we get:

2
1 ∞ ∞  i  2π 2π  1 i j  1 k l 
SCS + ∆SCS = ∑∑   ∫0 dσ 1 ∫0 dσ 2 C  ψ 0ψ 0  ψ 0ψ 0  ×
2 n = 0 p = 0  2πα '  2  2 
× X a1 (σ 1 )... X a n (σ 1 ) X b1 (σ 2 )... X p (σ 2 ) × ∂ a1 ...∂ a n Fij ( x )∂ b1 ...∂ b p Fkl ( x ) B R . (147)
1 ~ ~ 1 ~ ~b
n! p!

~
Now we need to evaluate the 2-point functions of the X . The relevant contributions have non-
logarithmic finite parts and come from propagators for which there is no self-contraction. This
requires that n = p . Then we get a combinatorial factor of n! from the number of such
( )(
contractions in X (σ 1 ) X (σ 2 ) . The result is:
~ n ~ n
)
2
1 ∞ 1 i  2π 2π
SCS + ∆SCS = ∑   ∫ dσ 1 ∫ dσ 2 D a1b1 (σ 1 − σ 2 )...D a n bn (σ 1 − σ 2 ) ×
2 n = 0 n!  2πα '  0 0

1  1 
× ∂ a1 ...∂ a n Fij ( x )∂ b1 ...∂ bn Fkl ( x ) C  ψ 0iψ 0j  ψ 0kψ 0l  B . (148)
2  2  R

The fermion zero mode expectation values are evaluated using the recipe:

32
 ψ 0ψ 0 Fij → (− iα ')F
1 i j
(149)
2

where the F on the right hand side is a differential 2-form. Thus we are led to:

SCS + ∆SCS = T a1 ...a n ;b1 ...bn ∂ a1 ...∂ a n F ∧ ∂ b1 ...∂ bn F (150)

where

2
1 1 i  2 2π 2π
≡  (− iα ') ∫0 dσ 1 ∫0 dσ 2 D 1 1 (σ 1 − σ 2 )...D n n (σ 1 − σ 2 ) . (151)
a1 ...a n ; b1 ...bn ab a b
T 
2 n!  2πα ' 

Thence, we obtain:

SCS + ∆SCS = ∂ a1 ...∂ a n F ∧ ∂ b1 ...∂ bn F ×

2
1 1 i  2 2π 2π
×  (− iα ') ∫0 dσ 1 ∫0 dσ 2 D 1 1 (σ 1 − σ 2 )...D n n (σ 1 − σ 2 ) . (151b)
ab a b

2 n!  2πα ' 

Next we insert the expression for the propagator:

e −εm ab im (σ 2 −σ 1 )
( )
∞
D ab (σ 1 − σ 2 ) = α ' ∑ h e + hba e − im (σ 2 −σ 1 ) (152)
m =1 m

where ε is a regulator, and

1
h ij ≡ . (153)
g + 2πα ' B

As we have seen, this tensor when expanded about large B has the form:

h ij ≈
θ ij
−
(θgθ )ij + ... (154)
2πα ' 2πα '2

where the terms in the expansion are alternatively antisymmetric and symmetric, and the two
terms exhibited above are description-independent. Now we neglect all but the first term above.
It follows that

 ai bi ∞ e −εm + imσ e −εm −imσ

(α ')n ∫0 dσ
∞
11 2π n

T a1 ...a n ;b1 ...bn =
2π
∏ 

i =1 
h ∑ + h bi a i
∑  . (155)
2 n! m =1 m m =1 m 

After evaluating the sum over m , the result, depending on the regulator ε , is

11
(α ')n ∫0 dσ
2π n
(
 [ai bi ] 1 − e −ε + iσ ) 
∏ − h(ai bi ) ln 1 − e −ε + iσ
2
T a1 ...a n ;b1 ...bn =  − h  . (156)
2 n! 2π i =1 
ln
1− e (
− ε − iσ
) 

33
Here, h[ab ] and h (ab ) are, respectively, the antisymmetric and symmetric parts of h ab . The large- B
or Seiberg-Witten limit consists of the replacements:

θ ab
h[ab ] → , h (ab ) → 0 . (157)
2πα '

By virtue of the fact that

(1 − e − ε + iσ
) = i(σ − π )
lim ln
ε →0 (1 − e − ε − iσ
) (158)

this limit leads to an elementary integral. Evaluating it, one finally obtains the result:

1 (6 ) ∞
C ∧ ∑ (− 1) 2 j
1
SCS + ∆SCS = ∫ θ a1b1 ...θ 2 j 2 j ∂ a1 ...∂ a2 j F ∧ ∂ b1 ...∂ b2 j F =
j a b

2 j =0 2 (2 j + 1)!

= ∫ C (6 ) ∧ (F ∧ F )∗2 (159)
1
2

where the product ∗2 is defined by the following equation:

1s r 
sin  ∂ pθ pq ∂ q 
Fij ( x ), Fkl ( x ) ≡ Fij ( x )  s  F ( x ) . (160)
2
∗2 1 r kl
∂ pθ ∂ q
pq

This agree perfectly with eq. (144), the prediction from noncommutativity.
Now we extend the calculation from the point of eq. (156). To go to first order beyond the
Seiberg-Witten limit, we make the replacements:

θ ab θ ac g cdθ db
h[ab ] → , h (ab ) → − (161)
2πα ' (2πα ')2

and keep all terms that are first order in h (ab ) . Denote by T(1a)1 ...a n ;b1 ...bn the first correction to T
(defined in eq. (151)) away from the Seiberg-Witten limit. Then, we see that

T a1 ...a n ; b1 ...bn
=
(− i)
n −1
1 1 (θgθ ) 1 1 a2 b2 a n bn 2π dσ
θ ...θ ∫
ab
(σ − π )n −1 ln 1 − eiσ . (162)
2
(1)
2 (n − 1)! (2π ) 2πα '
n 0 2π

The integral in the above expression vanishes for even n . For odd n = 2 p + 1 , we find that

a ...a 2 p +1 ; b1 ...b2 p +1
T(1)1 =
(− 1)p 1 1 (θgθ ) 1 1 θ a2b2 ...θ a2 p+1b2 p+1 I
ab
(163)
2 (2 p )! (2π )2 p +1 2πα ' 2 p +1

where

34
 dσ
(σ − π )2 p ln 1 − eiσ = 2(− 1)p (2 p )!∑ (− 1) j π ζ (2 p − 2 j + 1) , (164)
p −1 2j
2π 2
I 2 p +1 ≡ ∫
0 2π j =0 (2 j + 1)!
where ζ is the Riemann’s zeta function.
Thence, we obtain:

T
a1 ...a 2 p +1 ; b1 ...b2 p +1
=
(− 1)
p
1 1 (θgθ ) 1 1 θ a2b2 ...θ a2 p+1b2 p+1 ×
ab

(1)
2 (2 p )! (2π )2 p +1 2πα '
dσ
(σ − π )2 p ln 1 − eiσ = 2(− 1)p (2 p )!∑ (− 1) j π ζ (2 p − 2 j + 1) . (164b)
p −1 2j
2π 2
×∫
0 2π j =0 (2 j + 1)!

It is convenient to define a 4-form W4 that encodes the derivative corrections for the coupling to
C (6 ) :
1
SCS + ∆SCS = ∫ C (6 ) ∧ F ∧ F + ∫ C (6 ) ∧ W4 . (165)
2

The leading-order term in W4 in the large- B limit is:

W4(0 ) = F ∧ F ∗2
− F ∧ F . (166)

The calculations leading to eq. (164) amount to computing W4 to first order (in α ' ) around the
Seiberg-Witten limit:

W4(1)
=∑
∞
1(θgθ )a b a b
θ ...θ
a
1 1
2 2 2 p +1b2 p +1
p −1
∂ a1 ...∂ a2 p +1 F ∧ ∂ b1 ...∂ b2 p +1 F × ∑ (− 1)
j π2j
ζ (2 p − 2 j + 1)
p =1 (2π )
2 p +1
2πα ' j =0 (2 j + 1)!
(167)

Interchanging the order of the two summations, we find that the sum over j can be performed
and leads to the appearance of the familiar ∗2 product. The result, after some relabeling of
indices, is:

W4 = ∑(1)
∞
1 (θgθ )cd ζ (2 p + 3)θ a b ...θ a 1 1 2 p + 2 b2 p+ 2
× ∂ c ∂ a1 ...∂ a2 p + 2 F ∧ ∂ d ∂ b1 ...∂ b2 p+ 2 F . (168)
p =0 (2π ) 2 p +3
2πα ' ∗2

Unlike the leading term eq. (159), which is a single infinite series in derivatives summarised by the
∗2 product, here we see a double infinite series. After forming the ∗2 product we still have an
additional series whose coefficients are ζ -functions (Riemann’s zeta functions) of odd argument.
We will now use this to extract the term corresponding to our computation in eq. (168) and
compare the two expressions. The computation is an evaluation of the amplitude:

1 3
∞ − ,−
Α 2 ≡ ∫ dy VRR2
−∞
2
(q; i )VO0 (a1, k1;0)VO0 (a2 , k2 ; y ) (169)

35
 1 3
− ,−  1 3
where V RR is the vertex operator for an RR potential of momentum q , in the  − ,− 
2 2

 2 2
0
picture, and VO are vertex operators for massless gauge fields of momentum ki and polarizations
ai , i = 1,2 . We define:

1 1
t ≡ α ' k1 ⋅ k2 = α ' k1iG ij k2 j ; a≡ k1 × k2 = k1iθ ij k2 j (170)
2π 2π

and change integration variables via y = − cot πτ . Then it follows that the coefficient of F ∧ F
provided by this computation, to be compared with the coefficient of eq. (168), is:

1
Γ(1 + 2t )
22t ∫ 2 dτ (cos πτ ) cos 2πaτ =
2t 1
. (171)
0 2 Γ(1 + a + t )Γ(1 − a + t )

The right hand side can be expanded in powers of t and, up to terms of Ο t 2 , one has: ( )
Γ(1 + 2t )
=
1
Γ(1 + a + t )Γ(1 − a + t ) Γ(1 − a )Γ(1 + a )
[
1 − (2γ + ψ (1 − a ) + ψ (1 + a ))t + Ο t 2 ( )] (172)

where γ is the Euler constant and ψ ( x ) is the digamma function ln Γ( x ) .

d
dx
The first term can be recognised as the kernel of the ∗2 -product, using the relation:

1 sin πa
= . (173)
Γ(1 − a )Γ(1 + a ) πa

Let us now examine the second term more carefully. We use the fact that:

∞
ψ (1 + x ) = −γ + ∑ (− 1) ζ (k )x k −1 (174)
k

k =2

to write:

( )
∞ ∞
2γ + ψ (1 − a ) + ψ (1 + a ) = −∑ 1 − (− 1) ζ (k )a k −1 = −2 ∑ ζ (2 p + 3)a 2 p + 2 . (175)
k

k =2 p=0

Putting everything together, we find that:

k ×k 
sin  1 2  ∞
W4(1) =  2 
k1 × k2 ∑
1
( ~
)
ζ (2 p + 3)(k1 × k 2 )2 p + 2 k1i (θgθ )ij k2 j F (k1 ) ∧ F (k2 ) . (176)
~
p = 0 2πα ' (2π )
2 p +3

On Fourier transforming, this is identical to eq. (168).

Thence, we obtain:
36
    
( )
∞ 1

1 −  − 2 ∑ ζ (2 p + 3)a t + Ο t  = 2 ∫ dτ (cos πτ ) cos 2πaτ , (176b)

1 1 2 p+2
  2 2t 2 2t

2 Γ(1 − a )Γ(1 + a )   p = 0 
 
0

1 (6 )
C ∧ F ∧ F + ∫ C (6 ) ∧
2∫
SCS + ∆SCS =

k ×k 
sin  1 2  ∞
∧ 
2 
∑
1
( )
ζ (2 p + 3)(k1 × k2 )2 p + 2 k1i (θgθ )ij k2 j F (k1 ) ∧ F (k2 ) , (176c)
~ ~
p = 0 2πα ' (2π )
k1 × k2 2 p+3

where ζ is the Riemann’s zeta function.

Now we consider Chern-Simons gauge theory with the gauge group G a complex Lie group such
as SL(2, C ) . Let Α be a connection on a G -bundle E over a three-manifold M . Such a
connection has a complex-valued Chern-Simons invariant

 
W (Α ) =
1 2
4π ∫
M
Tr  Α ∧ dΑ + Α ∧ Α ∧ A  , (177)
 3 

which we have normalized to be gauge-invariant modulo 2π . Just as in the case of a compact

group, W (Α ) is gauge-invariant modulo 2π . The indeterminacy in W (Α ) is real, even if the
gauge group is complex, because a complex Lie group is contractible to its maximal compact
subgroup. The quantum theory is based upon integrating the expression exp(iI ) , and for this
function to be well-defined, I must be defined mod 2π . Because of the indeterminacy in Re W ,
its coefficient must be an integer, while the coefficient of Im W may be an arbitrary complex
number. The action therefore has the general form

I = − s Im W + l ReW . s ∈ C , l ∈ Z . (178)

Alternatively, we can write

~ ~
tW t W t  2  t  2 
I=
2
+
2
=
8π ∫M 
Tr  Α ∧ d Α +
3
Α ∧ Α ∧ Α  +
 8π
∫ M
Tr  Α ∧ dΑ + Α ∧ Α ∧ Α  , (179)
 3 

with
~
t = l + is , t = l − is . (180)

Introduce a complex variable z and a complex-valued polynomial

n
g ( z ) = ∑ a j z j . (181)
j =0

Now consider the integral

37
 Z g = ∫ d 2 z exp(g ( z ) − g ( z )) . (182)

This again is a convergent oscillatory integral and a closer analogous of Chern-Simons theory,
with − ig ( z ) and ig ( z ) corresponding to the terms tW and ~
t W in the action (179).
Define a new polynomial
n
g~ ( z ) = ∑ a~j z j , (183)
j =0

and generalize the integral Z g to

Z g , g~ = ∫ d 2 z exp( g ( z ) − g~ ( z )) . (184)

Then Z g , g~ coincides with the original Z g if g~ = g .

Denoting the independent complex variables as z and ~z , the analytically continued integral is

z exp( g ( z ) − g~(~
Z g , g~ = ∫ dzd~ z )) . (185)
C

The integral is over a two-dimensional real integration cycle C , which is the real slice ~z = z if
g~ = g , and in general must be deformed as g and g~ vary so that the integral remains
convergent.
Just as we promoted z to a complex variable ~z that is independent of z , we will have to
~
promote Α to a new G -valued connection Α that is independent of Α . Thus we consider the
~
classical theory with independent G -valued connections Α and Α (we recall that G is a
complex Lie group such as SL(2, C ) ) and action

~
( ~
)
I Α, Α =
t
8π ∫M 
Tr

Α ∧ d Α +
2
Α ∧ Α ∧ Α
 t
 + ∫
~ ~ 2 ~ ~ ~
Tr  Α ∧ dΑ + Α ∧ Α ∧ Α  . (186)
3  8π M
 3 

Then we have to find an appropriate integration cycle C in the path integral

∫ DΑDΑ exp(iI (Α, Α)) .

~ ~
(187)
C

~
The cycle C must be equivalent to Α = Α if s is real, and in general must be obtained by
deforming that one as s varies so that the integral remains convergent.

Let us consider a noncommutative Euclidean Dp-brane with an even number p+1 of world-volume
directions. Given a collection of local operators Ο I ( x ) on the brane world-volume which
transform in the adjoint under gauge transformations, one can obtain a natural gauge-invariant
operator of fixed momentum k i by smearing the locations of these operators along a straight
contour given by ξ i (τ ) = θ ij k jτ with 0 ≤ τ ≤ 1 , and multiplying the product by a Wilson line
W ( x, C ) along the same contour,

38
  1 ∂ξ i (τ ) ˆ 
W ( x, C ) ≡ exp i ∫ dτ Ai ( x + ξ (τ )) , (188)
 0 ∂τ 

where  denotes the noncommutative gauge field, and F̂ the corresponding field strength. The
resulting formula for the gauge-invariant operator is:

p +1    
n n
Q (k ) = ∫ ( ) Ο I ( x + ξ (τ I )) ∗ eik . x =
1
∏ ∏
1
d x 
 ∫ d τ P
I  ∗ W x , C
(2π ) p +1
 I =1 0   I =1 
 n

d p +1 xL∗ W ( x, C )∏ Ο I ( x ) ∗ eik . x (189)
1
=∫
(2π ) p +1
 I =1 

where P∗ denotes path-ordering with respect to the ∗ -product, while L∗ is an abbreviation for the
combined path-ordering and integrations over τ I . In this formula the operators Ο I are smeared
over the straight contour of the Wilson line. This prescription arises by starting with the
symmetrised-trace action for infinitely many D-instantons and expanding it around the
configuration describing a noncommutative Dp-brane.
Expanding the Wilson line, we get

∞
Q (k ) = ∑ ∫ d p +1 xQm ( x )eik . x
1
(190)
m=0 (2π ) p +1

where
Qm (x ) = (θ∂ )i1 ...(θ∂ )im Ο1 (x ),..., Ο n (x ), Aˆi1 (x ),..., Aˆim (x )
1
. (191)
m! ∗m + n

Thence, we obtain the following expression:

∞
Q (k ) = ∑ ∫ (θ∂ )i1 ...(θ∂ )im Ο1 (x ),..., Ο n (x ), Aˆi1 (x ),..., Aˆim (x )
1 1
d p +1 x ⋅ eik . x . (191b)
m =0 (2π ) p +1
m! ∗m + n

Here we have introduced the notation f1 ( x ), f 2 ( x ),..., f p ( x ) ∗ p for the ∗ p product of p functions.
We note here the simple formula for ∗2 :

1 s r 
sin  ∂ pθ pq ∂ q 
f ( x ), g ( x ) ∗ ≡ f (x )  s  g ( x ) . (192)
2
1 r
∂ pθ ∂ q
2
pq

If the zero-momentum coupling is

Σ(0 )∫ Ο Σ ( A, X ) (193)
~ 1 1
d p +1 x
(2π ) p +1
Pfθ

39
where A is the gauge field and X are the transverse scalars, then the coupling at nonzero
momentum is given by

Σ(− k )∫ L∗ [Ο Σ ( A, X )W ( x, C )] ∗ eik . x . (194)

~ 1 1
d p +1 x
(2π ) p +1
Pfθ

The constant factor Pf θ ≡ det θ ij has been written explicitly, instead of absorbing it into the
definition of ΟΣ , for convenience. In our case the relevant closed string mode is the RR gauge
potential C p +1 (k ) , and, the role of ΟΣ is played by the operator µ p PfQ where µ p is the brane
~

tension and
Q ij ≡ θ ij − θ ik Fˆklθ lj . (195)

Hence we deduce the coupling of this brane to the form C ( p +1) , in momentum space, to be:

µ pε
i1 ...i p +1
Ci(1 p...+i 1p +)1 (− k )∫
~
(2π )
1
p +1 
( )
d p +1 xL∗  det 1 − θFˆ W ( x, C ) ∗ eik . x . (196)


On the other hand, we know that the coupling of a Dp-brane to a C ( p +1) form is given in the
commutative description by

Ci(1 p...+i p1+)1 (− k )δ ( p +1) (k ) . (197)

~
µ pε
i1 ...i p +1

As equations (196) and (197) describe the same system in two different descriptions, they must be
equal. Thus we predict the identity:

δ ( p +1) (k ) = ∫
1
(2π )p +1 
( 
)
d p +1 xL∗  det 1 − θFˆ W ( x, C ) ∗ eik . x . (198)

This amounts to saying that the right hand side is actually independent of  , a rather nontrivial
fact. Is interesting to express the eq. (198) in the operator formalism. We use the fact that

1 1
∫ d p +1 x → tr (199)
(2π )
p +1
2
Pfθ

to rewrite the left hand side of eq. (198) as follows:

δ p +1 (k ) = ∫
1
d p +1 xeik . x =
1
( )
tr Pfθeik . x . (200)
(2π ) p +1
(2π )
p +1
2

The right hand side of eq. (198) can be converted to a symmetrised trace involving
X i ≡ x i + θ ij Aˆ j ( x ) , and it becomes:

40
 1
p +1
(
Str PfQeik . X ) (201)
(2π ) 2

where Str denotes the symmetrised trace. Finally, we use x i , x j = iθ ij and X i , X [ ] [ j

] = iQ (x ) .
ij

Then eq. (198) takes the elegant form:

( [ ] ) ( [
tr Pf x i , x j eik . x = Str Pf X i , X j eik . X . (202) ] )
In this form, it is easy to see that eq. (198) holds for constant F̂ , or equivalently for constant Q . In
this special case it can be proved by pulling PfQ out of the symmetrised trace on the right hand
side, and then using eq. (199) with θ replaced by Q .
Now let us turn to the coupling of a noncommutative p-brane to the RR form C ( p −1) . In the
commutative case this form appears in a wedge product with the 2-form B + F . For the
noncommutative brane in a constant RR background, B + F must be replaced by the 2-form Q −1
with Q ij given by eq. (195). It follows that the coupling in the general noncommutative case (with
varying C ( p −1) ) is:

µ pε
i1 ...i p +1
Ci(1 p...−i 1p −)1 (− k )∫
~ 1
(2π ) p +1 
(
d p +1 xL∗  det 1 − θFˆ Q −1 )( ) i p i p +1 W ( x, C ) ∗ eik . x . (203)


For comparison, the coupling of a Dp-brane to the form C ( p −1) in terms of commutative variables
is given by:

µ pε
i1 ...i p +1
Ci(1 p...−i 1p −)1 (− k )∫
~
(2π )
1
p +1
d p +1 x(B + F )i p i p +1 ( x )eik . x = µ pε 1
i ...i p +1 ~
[ ~
]
Ci(1 p...−i 1p −)1 (− k ) δ p +1 (k )Bi p i p +1 + Fi p i p +1 (k ) .

(204)
−1
Next, rewrite Q as

Q −1 = θ −1 1 + θFˆ 1 − θFˆ

( )
−1
(
 = B + Fˆ 1 − θFˆ

)−1
(205)

where we have used the relation B = θ −1 . Using this relation and also eq . (198), we can rewrite eq.
(203) as:

µ pε
i1 ...i p +1 
Ci(1 p...−i 1p −)1 (− k )δ p +1 (k )Bi p i p +1 + ∫
~ 1
(2π ) p +1 d
p +1 
(
xL∗  det 1 − θFˆ  Fˆ 1 − θFˆ

) ( )
−1
 W ( x, C ) ∗ eik . x  .
i p i p +1  
   
(206)
Equating this to eq. (204), we find that

Fij (k ) = ∫
~
(2π )
1
p +1

(
d p +1 xL∗  det 1 − θFˆ  Fˆ 1 − θFˆ

) ( ) −1
 W ( x, C ) ∗ eik . x . (207)
ij 
 

This relates the commutative field strength F to the non-commutative field strength F̂ ,
therefore it amounts to a closed-form expression for the Seiberg-Witten map.
41
The coupling of a noncommutative Dp-brane to the RR form C ( p − 3) for the case of constant RR
field is:
1
2
µ pε 1 p +1 Ci(1 p...−i p3−)3 (0)∫
i ...i ~ 1
(2π ) p +1
( )
d p +1 x det 1 − θFˆ (Q −1 )i p − 2i p−1 (Q −1 )i p i p +1 (208)

where the 2-form Q −1 is given in eq. (205). For spatially varying C ( p − 3) we can therefore write the
coupling as:

1
2
µ pε 1 p+1 Ci(1 p...−i p3−)3 (− k )∫
i ...i ~ 1
(2π ) p +1
(
d p +1 xL∗[ det 1 − θFˆ Q −1 )( ) i p − 2 i p −1 (Q ) −1
i p i p +1 W ( x, C )] ∗ eik . x . (209)

In the DBI approximation of slowly varying fields, the commutative coupling is:

µ pε 1 p+1 Ci(1 p...−i p3−)3 (− k )∫ d p +1 x(B + F )i p − 2i p−1 (B + F )i p i p+1 eik . x . (210)

1 i ...i ~ 1
2 (2π ) p +1

Inserting eq. (205) for Q −1 in eq. (209), and comparing with eq. (210), we find that in the DBI
approximation we must have:

∫ d k ' Fij (k ')Fkl (k − k ') = ∫

p +1 ~
(2π )
1
p +1

( ) (
d p +1 xL∗  det 1 − θFˆ  Fˆ 1 − θFˆ

)
−1
(
  Fˆ 1 − θFˆ
ij 
)−1
 W ( x, C ) ∗ eik . x
 kl 
 
(211)
To arrive at this expression we have made use of the identities eqs. (198) and (207).
The open Wilson line including transverse scalars is given by:

 1  ∂ξ i (τ ) ˆ 
W ' ( x, C ) = P∗ exp i ∫ dτ  ˆ a ( x + ξ (τ )) . (212)
Ai ( x + ξ (τ )) + qa Φ 

0
 ∂τ 

Inserting this definition in place of W ( x, C ) in eq. (190), and denoting the left hand side by Q ' (k ) ,
one finds that the couplings to a general spatially varying supergravity mode are:

∞
Q ' (k ) = ∑ ∫ d p +1 xQm' ( x )eik . x
1
(213)
m =0 (2π ) p +1

where the Qm' are given by

1 m  m
Qm' ( x ) = ∑  (θ∂ ) 1 ...(θ∂ ) k (iq )a k +1 ...(iq )am Ο1 ( x ),.., Ο n ( x ), Aˆi1 ( x ),.., Aˆik ( x ), Φ
ˆ a k +1 ( x ),.., Φ
ˆ am (x )
i i
.
m! k = 0  k  ∗n+ m

(214)

Thence, we obtain that:

42
 ∞
Q ' (k ) = ∑ ∫
1
d p +1 x ×
m =0 (2π ) p +1

m
 m
 (θ∂ ) 1 ...(θ∂ ) k (iq )a k +1 ...(iq )a m Ο1 ( x ),.., Ο n ( x ), Aˆi1 ( x ),.., Aˆik ( x ), Φ
ˆ a k +1 ( x ),.., Φ
ˆ am (x )
1
× ∑
i i
eik . x .
m! k = 0  k  ∗ n +m

(214b)
It is well-known that non-BPS branes in superstring theory also couple to RR forms. In
commutative variables, these couplings for a single non-BPS brane are given by:

1
SˆCS = µ p −1 ∫ dT ∧ ∑ C (n ) ∧ e B + F (215)
2T0 n

where T is the tachyon field and T0 is its value at the minimum of the tachyon potential.
Consider the coupling of a Euclidean non-BPS Dp-brane with an even number of world-volume
directions, to the RR form C ( p ) , in the commutative description:

µ p −1 ∫ d p +1 xε 1 2 p +1 ∂ i1 T ( x )Ci(2p...)i p +1 ( x ) =
1 1
µ p −1 ∫ dT ∧ C ( p ) =
i i ...i

2T0 2T0

=
1
2T0
i ...i
( ~ ~
)
µ p −1 ∫ d p +1kε 1 p +1 − iki1 T (k )Ci(2p...)i p +1 (− k ) . (216)

The noncommutative generalisation of this coupling, for constant RR fields, is

1
2T0
µ p −1ε 1 2 p +1 Ci(2p...)i p+1 (0)∫
i i ...i ~ 1
(2π ) p +1
(
d p +1 x det 1 − θFˆ Di1 Tˆ ( x ) (217) )
where
[
DiTˆ ( x ) = −iQij−1 X j , Tˆ ( x ) . (218) ]
Then, the same RR form C ( p ) couples to a noncommutative non-BPS Dp-brane through the
following coupling for each momentum mode:

1
2T0
µ p −1ε 1 2 p +1 Ci(2p...)i p +1 (− k )∫
i i ...i ~ 1
(2π ) p +1 
( 
)
d p +1 xL∗  det 1 − θFˆ Di1 Tˆ ( x )W ( x, C ) ∗ eik . x . (219)

4. On some equations concerning the noncommutative quantum mechanics regarding the

particle in a constant field and the noncommutative classical dynamics related to quadratic
Lagrangians [8]

We consider here the most simple and usual noncommutative quantum mechanics (NCQM) which
is based on the following algebra:

[xˆ , pˆ ] = ihδ
k j kj , [xˆ , xˆ ] = ihθ
k j kj , [pˆ , pˆ ] = 0
k j (220)

43
where Θ = (θ kj ) is the antisymmetric matrix with constant elements. To find elements Ψ ( x, t ) of
the Hilbert space in ordinary quantum mechanics (OQM), it is usually used the Schrodinger
equation
∂
ih Ψ ( x, t ) = Hˆ Ψ ( x, t ) , (221)
∂t

which realizes the eigenvalue problem for the corresponding Hamiltonian operator
  ∂
Hˆ = H ( pˆ , x, t ) , where pˆ k = −ih
 .
  ∂xk
There is another approach based on the Feynman path integral method

( x '' ) i 
Κ ( x ' ' , t ' ' ; x ' , t ') = ∫ exp S [q ]Dq , (222)
( x')  h 

where Κ ( x' ' , t ' ' ; x' , t ') is the kernel of the unitary evolution operator U (t ) acting on Ψ ( x, t ) in
( )
L2 R D .
Path integral in its most general formulation contains integration over paths in the phase space
R 2 D = {( p, q )} with fixed end points x' and x' ' , and no restrictions on the initial and final values of
the momenta, i.e.

( x '' )  2πi t ''

Κ ( x ' ' , t ' ' ; x ' , t ') = ∫
( x')
exp ∫ [ pk q&k − H ( p, q, t )]dt DqDp , (223)
 h t '


The Feynman path integral for quadratic Lagrangians can be evaluated analytically and the exact
expression for the probability amplitude is:

 ∂ 2S   2πi 
Κ ( x ' ' , t ' ' ; x ' , t ') = S ( x' ' , t ' ' ; x' , t ') , (224)
1
det  − '' '  × exp
 ∂x ∂x 
(ih ) 2
D
 k j   h 

where S ( x' ' , t ' ' ; x' , t ') is the action for the classical trajectory which is the solution of the Euler-
Lagrange equation of motion.
Thence, we obtain that:

( x '' )  2πi t '' 

Κ ( x ' ' , t ' ' ; x ' , t ') = ∫
( x')
exp ∫ [ pk q& k − H ( p, q, t )]dt DqDp =
 h t' 
 ∂2S   2πi 
S ( x' ' , t ' ' ; x' , t ') . (224b)
1
= det  − '' '  × exp
 ∂x ∂x 
(ih ) 2
D
 k j   h 

If we known the following Lagrangian

L( x&, x, t ) = αx& , x& + β x, x& + γx, x + δ , x& + η , x + φ , (225)

and algebra (220), we can obtain the corresponding effective Lagrangian

44
 Lθ (q& , q, t ) = αθ q& , q& + βθ q, q& + γ θ q, q + δθ , q& + ηθ , q + φθ (226)

that is suitable for quantization with path integral in NCQM. Exploiting the Euler-Lagrange
equations
∂Lθ d ∂Lθ
− = 0 , k = 1,2,..., D
∂qk dt ∂q&k

one can obtain classical path qk = qk (t ) connecting given end points x' = q(t ') and x' ' = q(t ' ') . For
this classical trajectory one can calculate action

Sθ ( x' ' , t ' ' ; x' , t ') = ∫ Lθ (q& , q, t )dt .

t ''

t'

Path integral in NCQM is a direct analog of (222) and its exact expression in the form of quadratic
actions Sθ ( x' ' , t ' ' ; x' , t ') is

 ∂ 2 Sθ   2πi 
Κ θ ( x ' ' , t ' ' ; x ' , t ') = Sθ ( x' ' , t ' ' ; x' , t ') . (227)
1
det  − '' '  × exp
 ∂x ∂x 
(ih ) 2
D
 k j   h 

For a particle in a constant field, the Lagrangian on commutative configuration space is:

L( x&, x ) =
m 2
2
( )
x&1 + x&22 − η1 x1 − η 2 x2 . (228)

The corresponding data in the matrix form are:

I , β = 0 , γ = 0 , δ = 0 , η τ = (− η1 ,−η 2 ) , φ = 0 , (229)
m
α=
2

where I is 2 × 2 unit matrix. We now note that one can easily find

mθ
(− η2 ,η1 ) , φθ = mθ η12 + η22 . (230) ( )
2
m
αθ = I , βθ = 0 , γ θ = 0 , ηθ = η , δθτ =
2 2 8

In this case, it is easy to find the classical action. The Lagrangian Lθ (q& , q, t ) is

Lθ =
1
2
( 1
2
)
m q&12 + q&22 + mθ (η1q&2 − η2 q&1 ) − η1q1 − η 2 q2 + mθ 2 η12 + η 22 . (231)
1
8
( )
The Lagrangian given by (231) implies the Euler-Lagrange equations

mq&&1 = −η1 , mq&&2 = −η2 . (232)

Their solutions are:

45
 η1t 2 η 2t 2
q1 (t ) = − + tC2 + C1 , q2 (t ) = − + tD2 + D1 , (233)
2m 2m

where C1 , C2 , D1 and D2 are constants which have to be determined from conditions:

q1 (0 ) = x1' , q1 (T ) = x1'' , q2 (0 ) = x2' , q2 (T ) = x2'' . (234)

After finding the corresponding constants, we have

η jt 2 ηT ηt 1 ηT
q j (t ) = x −
'
j
1
( )
+ t  x 'j' − x 'j + j  , q& j (t ) = − j + x 'j' − x 'j + j , ( ) j = 1,2 . (235)
2m T 2m  m T 2m

Using (234) and (235), we finally calculate the corresponding action

[( ) ( ) ]− 12 T [η (x ) ( )]
T
Sθ ( x' ' , T ; x' ,0) = ∫ Lθ (q& , q, t )dt =
1
+ x1' + η 2 x2'' + x2' +
2 2
m x1'' − x1' + x2'' − x2' 1
''
1
0
2T
1
[ ( )
+ mθ η1 x2'' − x2' − η 2 x1'' − x1' −
2
(
1 3 2
24m
1
)] (
T η1 + η22 + mθ 2T η12 + η 22 . (236)
8
) ( )
According to (227) one gets

 2πi   2πi mθ
Κ θ ( x ' ' , T ; x ' ,0 ) = Sθ ( x' ' , T ; x' ,0 ) = Κ 0 ( x' ' , T ; x' ,0 )exp
1 m
exp ⋅
ih T  h   h 2

( ) ( 1
)
⋅ η1 x2'' − x2' − η 2 x1'' − x1' + θT η12 + η 22 ( )  , (237)
 4 

where Κ 0 ( x' ' , T ; x' ,0 ) is related to the Lagrangian (228) for which θ = 0 . Hence, in this case there
is a difference only in the phase factor. It is easy to see that the following connection holds:

 
Κ θ ( x' ' , T ; x' ,0 ) = Κ 0  x' '+ θTJη , T ; x' ,0  , (238)
1
 2 

where
 0 1
J =   .
 −1 0

Now taking N → ∞ one can rewrite the following equation

 2πiε
N +1 D

Κ ( x' ' , t ' ' ; x' , t ') = lim ∫∏

 2 
  det (α n ) × exp [ α n q&n , q&n + β n qn , q&n +
N →∞
R DN n =1  ihε   h

+ γ n qn , qn + δ n , q&n + ηn , qn + φn ])∏ d D qn , (239)

N

n =1

46
as (222):

 2πi  N   2 D N  2 
D

Κ ( x' ' , t ' ' ; x' , t ') = ∫ exp L(q& , q, t )dt  lim ∏   det (α (t ' ')) ×   det (α n )dqn .
t ''
 
 h t'
∫  N → ∞ n =1   ihε   ihε 

(240)

5. Mathematical connections [9]

Ramanujan’s modular equations and Palumbo-Nardelli model

Now, we note that the number 8, and thence the numbers 64 = 82 and 32 = 2 2 × 8 , are connected
with the “modes” that correspond to the physical vibrations of a superstring by the following
Ramanujan function:

 ∞ cos πtxw' 
− πx 2 w '
 ∫0 cosh πx e dx 
142
4 anti log πt 2 ⋅ 2
− t w'
 e 4 φw' (itw') 
w'
1 
8= . (241)
3   10 + 11 2   10 + 7 2  
log   + 
 


  4   4 


Furthermore, with regard the number 24 (12 = 24 / 2 and 32 = 24 + 8) they are related to the
physical vibrations of the bosonic strings by the following Ramanujan function:

 ∞ cos πtxw' 
− πx 2 w '
 ∫0 cosh πx
e dx 
142
4 anti log πt 2 ⋅ 2
 e
−
4
w'
φ (itw ' )  t w'
24 =  
w '
. (242)
  10 + 11 2   10 + 7 2  
log   + 
 


  4   4 


Palumbo (2001) ha proposed a simple model of the birth and of the evolution of the Universe.
Palumbo and Nardelli (2005) have compared this model with the theory of the strings, and
translated it in terms of the latter obtaining:

− g µρ g νσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =

 R 1 1 
− ∫ d 26 x g −
 16πG 8 2 

( )
∞
 1 ~ 2 κ 2
2 
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  ,
1 1/ 2
(243)
0
2κ 10  2 g10 

47
A general relationship that links bosonic and fermionic strings acting in all natural systems.

p-adic, adelic and zeta-strings

Like in the ordinary string theory, the starting point of p-adic strings is a construction of the
corresponding scattering amplitudes. Recall that the ordinary crossing symmetric Veneziano
amplitude can be presented in the following forms:

 Γ(a )Γ(b ) Γ(b )Γ(c ) Γ(c )Γ(a )  ζ (1 − a ) ζ (1 − b ) ζ (1 − c )

A∞ (a, b ) = g 2 ∫ x ∞ 1 − x ∞ dx = g 2 
a −1 b −1
+ +  = g2 =
R
 Γ(a + b ) Γ(b + c ) Γ(c + a )  ζ (a ) ζ (b ) ζ (c )
 i µ
( )
4
= g 2 ∫ DX exp −
 2π ∫ d 2
σ∂ α
X ∂
µ α X ∏
 j =1
∫ d 2σ j exp ik µ( j ) X µ , (244)

where h = 1 , T = 1 / π , and a = −α (s ) = −1 − , b = −α (t ) , c = −α (u ) with the condition

s
2
s + t + u = −8 , i.e. a + b + c = 1 .
The p-adic generalization of the above expression

A∞ (a, b ) = g 2 ∫ x ∞ 1 − x ∞ dx ,
a −1 b −1
R

is:
Ap (a, b ) = g 2p ∫ x p 1 − x p dx , (245)
a −1 b −1
Qp

where ... p denotes p-adic absolute value. In this case only string world-sheet parameter x is
treated as p-adic variable, and all other quantities have their usual (real) valuation.
Now, we remember that the Gauss integrals satisfy adelic product formula

∫ χ ∞ (ax 2 + bx )d ∞ x∏ ∫ χ p (ax 2 + bx )d p x = 1 , a ∈ Q× , b ∈ Q , (246)

R Qp
p∈P

what follows from

χ v (ax 2 + bx )d v x = λv (a ) 2a v 2 χ v  −
−
1
 b2 
∫Qv
 4 a
 , v = ∞,2,..., p... . (247)


These Gauss integrals apply in evaluation of the Feynman path integrals

 1 t '' 
K v ( x ' ' , t ' ' ; x ' , t ') = ∫ L(q& , q, t )dt  Dv q , (248)
x ' ', t ''
χv  −∫
x ', t '
 h t '


for kernels K v ( x' ' , t ' ' ; x' , t ') of the evolution operator in adelic quantum mechanics for quadratic
Lagrangians. In the case of Lagrangian

48
 1  q& 2 
L(q& , q ) =  − − λq + 1 ,
2 4 

for the de Sitter cosmological model one obtains

K ∞ ( x' ' , T ; x' ,0)∏ K p ( x' ' , T ; x' ,0) = 1 , x' ' , x' , λ ∈ Q , T ∈ Q× , (249)
p ∈P

where
 λ2T 3 T ( x' '− x') 
2
+ [λ ( x' '+ x') − 2] +
1
K v ( x' ' , T ; x' ,0 ) = λv (− 8T ) 4T v 2 χ v  −
−
 . (250)
 24 4 8T 

Also here we have the number 24 that correspond to the Ramanujan function that has 24
“modes”, i.e., the physical vibrations of a bosonic string. Hence, we obtain the following
mathematical connection:
 λ2T 3 T ( x' '− x') 
2
+ [λ ( x' '+ x') − 2] +
1
K v ( x' ' , T ; x' ,0 ) = λv (− 8T ) 4T v χ v  −
−
2  ⇒
24 4 8 T 
 
 ∞ cos πtxw' 
− πx 2 w '
 ∫0 cosh πx e dx  142
4 anti log πt 2 ⋅ 2
− t w'
 e 4 φw ' (itw') 
w'

⇒  . (250b)
  10 + 11 2   10 + 7 2  
log   + 
 


  4   4 


The adelic wave function for the simplest ground state has the form

ψ ( x ), x ∈ Z
( )
ψ A ( x ) = ψ ∞ ( x )∏ Ω x p =  ∞ , (251)
p∈P 0, x ∈ Q \ Z

( ) ( )
where Ω x p = 1 if x p ≤ 1 and Ω x p = 0 if x p > 1 . Since this wave function is non-zero only in
integer points it can be interpreted as discreteness of the space due to p-adic effects in adelic
approach. The Gel’fand-Graev-Tate gamma and beta functions are:

ζ (1 − a ) 1 − p a −1
Γ∞ (a ) = ∫ x ∞ χ ∞ ( x )d ∞ x = , Γp (a ) = ∫ x p χ p ( x )d p x =
a −1 a −1
, (252)
R ζ (a ) Q p 1 − p−a
B∞ (a, b ) = ∫ x ∞ 1 − x ∞ d ∞ x = Γ∞ (a )Γ∞ (b )Γ∞ (c ) , (253)
a −1 b −1
R

B p (a, b ) = ∫ x p 1 − x p d p x = Γp (a )Γp (b )Γp (c ) , (254)

a −1 b −1
Qp

where a, b, c ∈ C with condition a + b + c = 1 and ζ (a ) is the Riemann zeta function. With a

regularization of the product of p-adic gamma functions one has adelic products:

49
 Γ∞ (u )∏ Γp (u ) = 1 , B∞ (a, b )∏ B p (a, b ) = 1 , u ≠ 0,1, u = a, b, c, (255)
p∈P p∈P

where a + b + c = 1 . We note that B∞ (a, b ) and B p (a, b ) are the crossing symmetric standard and
p-adic Veneziano amplitudes for scattering of two open tachyon strings. Introducing real, p-adic
and adelic zeta functions as

ζ ∞ (a ) = ∫ exp(− πx 2 ) x ∞ d ∞ x = π 2 Γ  , (256)
a
a −1 − a
R
2
ζ p (a ) =
1− p
1
−1 ∫Q
p
( ) a −1
Ω x p x p dpx =
1
1 − p−a
, Re a > 1 , (257)

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) , (258)
p ∈P

one obtains

ζ A (1 − a ) = ζ A (a ) , (259)

where ζ A (a ) can be called adelic zeta function. We have also that

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) = ∫ exp(− πx ) x
R
2 a −1
∞
d∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p d p x . (259b)
p ∈P p

( ) ( )
Let us note that exp − πx 2 and Ω x p are analogous functions in real and p-adic cases. Adelic
harmonic oscillator has connection with the Riemann zeta function. The simplest vacuum state of
the adelic harmonic oscillator is the following Schwartz-Bruhat function:

∏ Ω( x ),
1
ψ A (x ) = 2 e −πx
2
∞
4
p p (260)
p ∈P

whose the Fourier transform

∏ Ω (k )
1
ψ A (k ) = ∫ χ A (kx )ψ A (x ) = 2 4 e −πk
2
∞
p p (261)
p ∈P

has the same form as ψ A ( x ) . The Mellin transform of ψ A ( x ) is

( )
a
a −
Φ A (a ) = ∫ψ A ( x ) x d A× x = ∫ ψ ∞ ( x ) x Ω x p x d p x = 2Γ π 2 ζ (a ) (262)
1
d ∞ x∏
a −1 a −1
−1 ∫Q
a

p∈P 1 − p 2
R p

and the same for ψ A (k ) . Then according to the Tate formula one obtains (259).
The exact tree-level Lagrangian for effective scalar field ϕ which describes open p-adic string
tachyon is

50
 1 p2  1 − 2 


1
Lp =  − ϕp ϕ+ ϕ p +1  , (263)
g p −1  2
2
p +1 

where p is any prime number,
= −∂ t2 + ∇ 2 is the D-dimensional d’Alambertian and we adopt

metric with signature (− +... + ) . Now, we want to show a model which incorporates the p-adic
string Lagrangians in a restricted adelic way. Let us take the following Lagrangian

n −1 1  1 1 n +1 


−
L = ∑ Cn Ln = ∑ L = 2 
− φ ∑ n 2
φ +∑ φ  . (264)
n ≥1 n + 1
2 n
n ≥1 n ≥1 n g  2 n ≥1 

Recall that the Riemann zeta function is defined as

ζ (s ) = ∑
1 1
=∏ −s
, s = σ + iτ , σ > 1 . (265)
p 1− p
s
n ≥1 n

Employing usual expansion for the logarithmic function and definition (265) we can rewrite (264)
in the form
1 1 
 
L = − 2  φζ  φ + φ + ln (1 − φ ) , (266)
g 2  2  



where φ < 1 . ζ   acts as pseudodifferential operator in the following way:
2

ixk  k 2 ~

 r2
ζ  φ (x ) = ∫  2 φ (k )dk , − k = k0 − k > 2 + ε , (267)
1
e ζ  − 2 2

2 (2π )D

where φ (k ) = ∫ e(−ikx )φ ( x )dx is the Fourier transform of φ ( x ) .

~

Dynamics of this field φ is encoded in the (pseudo)differential form of the Riemann zeta function.
When the d’Alambertian is an argument of the Riemann zeta function we shall call such string
a “zeta string”. Consequently, the above φ is an open scalar zeta string. The equation of motion
for the zeta string φ is


  k2 ~ φ
∫k02 − kr 2 > 2 +ε e ζ  − 2 φ (k )dk = 1 − φ
1
ζ  φ = ixk
(268)
2 (2π )D

which has an evident solution φ = 0 .

For the case of time dependent spatially homogeneous solutions, we have the following equation
of motion
 − ∂2  − ik 0 t  k0  ~
2
φ (t )
ζ  t φ (t ) = φ (k0 )dk0 =
1
∫ e ζ  . (269)
 2  (2π ) k 0 > 2 +ε 
 2 1 − φ (t )

With regard the open and closed scalar zeta strings, the equations of motion are

51
 n ( n −1)

 ixk  k2 ~
ζ  φ =
1
ζ
∫  2 
e  − φ (k )dk = ∑ θ 2
φ n , (270)
2 (2π )D n ≥1

 n 2 n(n − 1) n (n2−1) −1 n +1 


ζ  θ =
1 ixk 
D ∫e ζ −

k2 ~
θ (k )dk = ∑ θ +
2(n + 1)
θ φ − 1  , (271) ( )
4 (2π )  4 n ≥1  

and one can easily see trivial solution φ = θ = 0 .

Now we take the eq. (87b) of Section 2. We note that are possible the following mathematical
connections with the Palumbo-Nardelli model (243) and the zeta strings (268):

Γ( N / 2 )
(2R π ) N −3 N − 3
Γ  q'
 2  1
V (r ) =
2π N / 2 ( N − 2) ε0 ∫
0
R N −2
dq ' ⇒

− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
 1 ~ 2 κ 2
2 
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  , (272)
1 1/ 2

0
2κ 10  2 g10 

Γ( N / 2 )
(2R π ) N −3  N − 3
Γ  q'
 2  1
V (r ) =
2π N / 2 ( N − 2) ε0 ∫
0
R N −2
dq ' ⇒

 k2 ~ φ
eixk ζ  − φ (k )dk =
1
(2π )D ∫
⇒ r . (273)
k 02 − k 2 > 2 + ε
 2 1−φ

Now, we take the right hand side of eq. (100c) of Section 3. We have the following mathematical
connections with the eqs. (129), (137) and (138) and with Palumbo-Nardelli model:

det (G + 2πα ' Φ )  1

SˆDBI SW =∫ 1 + θ ij ∂ j Aˆi + θ ijθ kl ∂ j Aˆ k , ∂ l Aˆi +
Gs  2 ∗2

1 1 
+ θ baθ kl Aˆ k , ∂ l Fˆab
+ θ ijθ kl Fˆ ji , Fˆlk  ⇒
2 ∗2 8 ∗2 

d
⇒≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA ⇒
C dt S
− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
1 / 2 −2Φ  1 ~ 2 κ 102 2 
=∫
1
2 ∫
d 10
x (− G ) e  R + 4∂ µ Φ ∂ µ
Φ − H 3 − 2 Trν F2  , (274)
0
2κ 10  2 g10 

52
 SˆDBI =
1
Gs ∫ 
( 
(
L∗  det G + 2πα ' Fˆ + Φ W ( x, C ) ∗ eik . x = ))
=
1
∫
 1 2  
( ) (
det G L∗ 1 − (2πα ') trG −1 Fˆ + Φ G −1 Fˆ + Φ W ( x, C ) ∗ eik . x + ... ⇒ )
Gs  4  
d
⇒≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA ⇒
C dt S
− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
 1 ~ 2 κ 102 2 
∫ d x(− G )
1 −2Φ µ
=∫  R + 4∂ µ Φ∂ Φ − H 3 − 2 Trν F2  , (275)
10 1/ 2
e
0
2κ 102  2 g10 

 2πα '  
det (G + 2πα ' Φ )  L∗ (W ( x, C )) +
1 1
SˆDBI = ∫ trMF + M θ ∂ j Flk , Ai ∗2 + trMF , trθF +
kl ij
∗2
Gs  2  2 

−
(2πα ')2 tr MF , MF + (2πα ')2 trMF , trMF + ... ⇒
∗2 ∗2 
4 8 
d
⇒≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA ⇒
C dt S
− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
1 / 2 −2Φ  1 ~ 2 κ 102 2 
=∫
1
2 ∫
d 10
x (− G ) e  R + 4∂ µ Φ ∂ µ
Φ − H 3 − 2 Trν F2  , (276)
0
2κ 10  2 g10 

thence, the mathematical connections between noncommutative DBI actions, Maxwell’s

equations and Palumbo-Nardelli model.

Now we take the eqs. (164b) e (176c) of Section 3. We note that are possible the following
mathematical connections with the eq. (268) concerning the zeta strings:

a ...a 2 p +1 ; b1 ...b2 p +1
T(1)1 =
(− 1)p 1 1 (θgθ ) 1 1 θ a2b2 ...θ a2 p+1b2 p+1 ×
ab

2 (2 p )! (2π )2 p +1 2πα '

dσ
(σ − π )2 p ln 1 − eiσ = 2(− 1)p (2 p )!∑ (− 1) j π ζ (2 p − 2 j + 1) ⇒
p −1 2j
2π 2
×∫
0 2π j =0 (2 j + 1)!
 k2 ~ φ
∫k 02 − kr 2 > 2+ε e ζ  − 2 φ (k )dk = 1 − φ , (277)
1
⇒ ixk

(2π )D

53
 1 (6 )
SCS + ∆SCS = ∫ C ∧ F ∧ F + ∫ C (6 ) ∧
2
k ×k 
sin  1 2  ∞
∧ 
2 
k1 × k2 ∑
1
( )
ζ (2 p + 3)(k1 × k2 )2 p + 2 k1i (θgθ )ij k2 j F (k1 ) ∧ F (k2 ) ⇒
~ ~
p = 0 2πα ' (2π )
2 p +3

2
ixk  k2 ~ φ
φ (k )dk =
1
⇒ D ∫k 2 − k 2 > 2 + ε
e ζ 
 − . (278)
(2π ) 0
r
 2 1−φ

Now we take the eqs. (211) and (219) of Section 3. Also these equations can be connected with
the Maxwell’s equations and with Palumbo-Nardelli model:

∫ d k ' Fij (k ')Fkl (k − k ') = ∫

p +1 ~
(2π )
1 
p +1

( ) (  
−1
)
( 
)
d p +1 xL∗  det 1 − θFˆ  Fˆ 1 − θFˆ   Fˆ 1 − θFˆ  W ( x, C ) ∗ eik . x
−1

 ij kl 
d
⇒≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA ⇒
C dt S
− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
1 / 2 −2Φ  1 ~ 2 κ 102 2 
=∫
1
2 ∫
d 10
x (− G ) e  R + 4∂ µ Φ ∂ µ
Φ − H 3 − 2 Trν F2  , (279)
0
2κ 10  2 g10 

1
2T0
µ p −1ε 1 2 p +1 Ci(2p...)i p +1 (− k )∫
i i ...i ~ 1
(2π )p +1 
( )
d p +1 xL∗  det 1 − θFˆ Di1 Tˆ ( x )W ( x, C ) ∗ eik . x ⇒

d
⇒≈ ∫ B ⋅ dl = µ0 I enc + µ0ε 0 ∫ E ⋅ nˆdA ⇒
C dt S
− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
 κ102 2 
x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3
1 1 ~ 2
=∫ ∫d −
10 1/ 2
Trν F2  . (280)
0
2κ 2
10  2 g 2
10 

In conclusion, we take the eq. (236). We note that are possible the following mathematical
connections with eqs. (129), (138), i.e. the DBI noncommutative actions, eqs. (179), (186), i.e. the
Chern-Simons theory , the Ramanujan’s modular equations and the Palumbo-Nardelli model:

54
 [ ] [ ]
m (x1'' − x1' ) + (x2'' − x2' ) − T η1 (x1'' + x1' ) + η 2 (x2'' + x2' ) +
T
Sθ ( x' ' , T ; x' ,0) = ∫ Lθ (q& , q, t )dt =
1 2 2 1
0
2T 2
1
[ ( ) (
+ mθ η1 x2'' − x2' − η 2 x1'' − x1' −
2
1 3 2
24m
)] ( 1
)
T η1 + η22 + mθ 2T η12 + η 22 ⇒
8
( )
det (G + 2πα ' Φ )  1
⇒∫ 1 + θ ij ∂ j Aˆi + θ ijθ kl ∂ j Aˆ k , ∂ l Aˆi +
Gs  2 ∗2

1 1 
+ θ baθ kl Aˆ k , ∂ l Fˆab + θ ijθ kl Fˆ ji , Fˆlk  ⇒
2 ∗2 8 ∗2 
~
t  2  t ~ ~ 2 ~ ~ ~
⇒ ∫
8π M 
Tr  Α ∧ dΑ + Α ∧ Α ∧ Α  +
3  8π M 
∫ Tr  Α ∧ dΑ + Α ∧ Α ∧ Α  ⇒
3 
 ∞ cos πtxw' −πx 2 w' 
 ∫0 cosh πx e dx  142
4 anti log πt 2 ⋅ 2
 e
−
4
w'
φ (itw ' )  t w'
1  w' 
⇒ ⇒
3   10 + 11 2   10 + 7 2  
log   + 
 


  4   4 

− g µρ g νσ Tr (Gµν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 R 1 1 
⇒ − ∫ d 26 x g −
 16πG 8 2 

( )
∞
 1 ~ 2 κ 2
2 
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  . (281)
1 1/ 2

0
2κ 10  2 g10 

5 −1
Also in these expressions is very evident the link between π and φ = , i.e. the Aurea ratio,
2
by the simple formula

arccos φ = 0,2879π . (282)

Acknowledgments

The co-author Nardelli Michele would like to thank Prof. Branko Dragovich of Institute of
Physics of Belgrade (Serbia) for his availability and friendship.

55
References

[1] Walter Droscher, Jochem Hauser – “Spacetime physics and advanced propulsion concepts” –
American Institute of Aeronautics and Astronautics – 2006;

[1b] Scala Sabato – "Simmetrizzazione delle equazioni di Maxwell con l'introduzione del campo
gravitazionale, un'idea bizzarra?", http://itis.volta.alessandria.it/episteme/ep6/ep6-maxw.htm

[2] A. W. McDavid and C. D. McMullen – “Generalizing Cross Products and Maxwell’s Equations
to Universal Extra Dimensions” –

[3] Sunil Mukhi – “Noncommutativity in String Theory” – IAS, Princeton, July 1-12 2002.

[4] Sumit R. Das, Sunil Mukhi and Nemani V. Suryanarayana – “Derivative Corrections from
Noncommutativity” – arXiv:hep-th/0106024v3 – 27.06.2001.

[5] Sunil Mukhi and Nemani V. Suryanarayana – “Open-String Actions and Noncommutativity
Beyond the Large-B Limit” – arXiv:hep-th/0208203v1 – 28.08.2002.

[6] Edward Witten – “Analytic Continuation Of Chern-Simons Theory” – arXiv:1001.2933v2

[hep-th] 21.01.2010.

[7] Sunil Mukhi and Nemani V. Suryanarayana – “Gauge-Invariant Couplings of Noncommutative

Branes to Ramond-Ramond Backgrounds” – arXiv:hep-th/0104045v2 – 09.04.2001.

[8] Branko Dragovich, Zoran Rakic – “Path Integrals in Noncommutative Quantum Mechanics” –
arXiv:hep-th/0309204v2 – 26.09.2003.

[9] Branko Dragovich: “Zeta Strings” – arXiv:hep-th/0703008v1 – 1 Mar 2007.

[10] Branko Dragovich: “Zeta Nonlocal Scalar Fields” – arXiv:0804.4114v1 – [hep-th] –

25 Apr 2008.

Sites

Eng. Rosario Turco

http://mathbuildingblock.blogspot.com/

Dr Michele Nardelli (various papers on the string theory)

http://xoomer.virgilio.it/stringtheory/
http://nardelli.xoom.it/virgiliowizard/

CNR SOLAR
http://150.146.3.132/perl/user_eprints?userid=36

ERATOSTENE group
http://www.gruppoeratostene.com
56