On the Andrica and Cramer s Conjectures.

Mathematical connections between

Number Theory and some sectors of String Theory

Rosario Turco, Maria Colonnese, Michele Nardelli 1, 2

Dipartimento di Scienze della Terra

Universit degli Studi di Napoli Federico II, Largo S. Marcellino, 10
80138 Napoli, Italy
2

Dipartimento di Matematica ed Applicazioni R. Caccioppoli

Universit 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 we have described, in the Section 1, some mathematics concerning the Andrica s
conjecture. In the Section 2, we have described the Cramer Shank Conjecture. In the Section 3, we
have described some equations concerning the possible proof of the Cramer s conjecture and the
related differences between prime numbers, principally the Cramer s conjecture and Selberg s
theorem. In the Section 4, we have described some equations concerning the p-adic strings and the
zeta strings. In the Section 5, we have described some equations concerning the -deformation in
toroidal compactification for N = 2 gauge theory. In conclusion, in the Section 6, we have described
some possible mathematical connections between various sectors of string theory and number
theory.

1. The Andrica s Conjecture [1]

In this section we will show some mathematics related to the Andrica s conjecture:
pn

pn

using some our results on Legendre s conjecture ([2]).

Andrica s conjecture
Andrica s conjecture is so defined:
Andrica s Conjecture is a conjecture of Numbers Theory, concerning the gaps between two
successive prime numbers, formulated by romeno s mathematician Dorin Andrica in 1986. It
affirms that, for every couple of consecutive numbers pn and pn+1, we have:
1

pn
If we pose gn = pn+1

pn

pn, then the conjecture can be written as

g n < 2 pn + 1 .

.
Now we propose some mathematics concepts useful for a possible proof based on some
demonstrations concerning the Legendre s Conjecture, and on square roots of numbers included in
the numeric gap between a square and the successive one.
Legendre s Conjecture, by Adrien Marie Legendre, affirms that exists always a prime number
between n2 and (n+1)2. This conjecture is one of problems of Landau and, till now, it has not been
demonstrated .
Some observations about Legendre s conjectures are:
between n2 and (n+1)2 don t exists always a prime number, but at least two.
ERATOSTENE Group has developed it, see [1] and [2].
Difference between to perfect squares in the range I = [n2, (n+1)2]
To examine the connection between Legendre s conjecture and the Andrica s conjecture, we
must introduce some concepts.
Let I the closed range of integers definite as I = [n2, (n+1)2].
Let Dqp the difference between to consecutive perfect squares, in the range I.
Lemma 1.
The difference Dqp between two consecutive perfect squares, in a closed range of integers I = [n2,
(n+1)2] is always an odd number.
Proof.
Dqp = (n +1)2

n2 = 2n +1

Since for every n, Dqp = 2n + 1, then is always odd.

Example:
n = 2, valid for all the n natural numbers.
Dqp = 32

22 =9 - 4 = 5 = 2*2 +1

Lemma 2.
The number n of integer included in a closed range of integers I = [n2, (n+1)2] is an even number.
Proof.
For Lemma 1, since number of integer in I is:
N = Dqp +1 = 2n + 2 = 2(n + 1), then N is an even number.
Example:
If n = 2 N = 2(3) = 6 . Indeed the numbers included in gap I are : 4, 5, 6, 7, 8, 9; with 5 and 7
prime numbers.
Square roots of numbers in gap I = [n2, (n+1)2].
Lemma 3.
The difference Drq of square roots of two numbers, also not consecutive (prime and composite), in
a closed range of integers I = [n2, (n+1)2], excepts the number (n +1)2, is smaller than 1.
Proof.
At the extremes of the range of integer, Drq is :
Drq = n n = 0 if we consider at beginning of the interval the difference with itself or Drq = (n+1)
n = 1. Therefore Drq changes between 0 and 1.
Lemma 3 excludes the numbers (n + 1)2, because in the second case the difference doesn t give a
decimal part after the point. For this Lemma 3 is to check between n2 and (n +1)2 -1.
Since we think true the Lengendre s conjecture (see [2]), then between n2 and (n + 1)2 exists at
least a prime number and therefore an integer, so between the values 0 and 1 assumed by Drq exist
some values smaller than 1.
Obviously since we make reference at integer numbers in the range I; it is indifferent that they are
prime or composite. Therefore it is possible the applicability of Legendre s conjecture.

Example Square roots for n = 2.

4 = 2,00
5 = 2,23

with 5 prime number pn

6 = 2,44
7 = 2,64 with 7 prime number pn+1
3

8 = 2,82
9 = 3,00
Lemma 4.
In the range of integers I = [n2, (n+1)2] exist at least two prime numbers.
Proof.
The Bertrand s postulate, that is true, says that if n is an integer with n > 1, then there is always a
prime number such that n < p < 2n .
If we define a = n2 then the interval that we are considering is [a, a + 1 + 2 a]. Now for n > 3 the
term a + 1 + 2 a > 2a; therefore certainly is applicable the Betrand s postulate but we observe also
that this interval is bigger than of that used in Bertrand s postulate, therefore it increases the
probability to find at least a second prime number; in fact for Prime Number Theorem is:
((n 1) 2 )

(n 2 )

2n 1
1
ln((n 1) 2 )

(1)

Note: the intervals that we consider are the smaller critical intervals where we could risk don t find
the second prime number, but that the (1) guarantee. In the case of Andrica s conjecture, we think,
moreover, that the two consecutive prime numbers exists also a notable distances or notable gaps.

Example:
In the interval I with n =2 we have the two consecutive prime numbers 5 and 7.
7 - 5 = 2,64

2.23 = 0,41 < 1

From (1) results

((n+1)2)- (n2)

2,27

We note that this value is related with the aurea ratio by the following expression:

2,27

1
2

21 / 7

with

21 / 7

35 / 7

56 / 7

1
4,58359
2

2,29179

5 1
1,61803398... , i.e. the aurea ratio.
2

Lemma 5.
The difference of square roots of two consecutive prime numbers that are in a closed interval of
integers I = [n2, (n+1)2], except the number (n+1)2, is smaller of 1.
Proof.
The Lemma 5 is a consequence of Lemmas 3 and 4.
4

It is not still a proof of Andrica s conjecture; because the consecutive prime numbers could belong
to different square intervals.
Prime numbers in different square intervals.
Some prime numbers belong to successive square intervals, also being valid the Andrica s
conjecture, for example 113 and 127: the first is included in the interval between 102 and 112, the
second, 127, is included in the interval between 112 and 122. Really the square interval is always
possible individualize only one: for example it is between 102 and 122.

Lemma 6.
The difference of square roots of two numbers included in a closed interval of integers I = [n2,
(n+1)2] with k 1, except the number (n+k)2, is lowest of 1 provided that if k > 1 the difference
(n+k)2 n2 0 mod 3.

Proof.
Lemma 6 can be demonstrated with all previous Lemmas, marking also that k tends only to increase
certainly the interval of squares, Therefore the Lemma 6 is a generalization of Lemma 5. In
particular if k = 1 we return at Lemma 5 and it doesn t occur to consider if the difference is a
multiple of 3.
For example 131 and 137 are two prime numbers and their difference is multiple of 3, but it doesn t
count because the interval is the same for both the prime numbers. In fact is [112, 122] with k = 1.
Instead if we look 113 and 137 the interval to consider is different. That is k = 2, in fact is [102, 122]
but 137 113 = 24 that is multiple of 3.
Here the difference between the square roots is greater than 1 when the difference is even and
multiple of 3 (it is the same to say that it is multiple of 6). But there is to say that 137 and 113 are
not neither consecutive prime numbers. The problem that the difference between square roots of
two consecutive prime numbers can be greater than 1 could be when the two square intervals are
not adjacent, thence for example for k > 2.
With regard the prime numbers 113 and 137 and their difference, i.e. 24, we note that it is possible
the following mathematical connection with the Ramanujan s modular function concerning the
physical vibrations of the bosonic strings:

4 anti log

cos txw'
e
cosh x
e

137 113 24
log

t2
w'
4
w'

x 2 w'

dx

142
t 2 w'

itw'

10 11 2
4

10 7 2
4

Lemma 7.
Two consecutive prime numbers, included in closed interval of integers I = [n2, (n+1)2] with k > 2,
haven t got a difference D = 6j when j > 1, or even and multiple of 3.
Proof.
If the difference of two consecutive prime numbers is even and multiple of 3, it could be:
D = 3m = 6j where m = 2j (even)
If j = 1 we have the situation k = 2 and of the prime numbers as 23 and 29 where D = 6 (j=1) and
the difference of square roots is smaller than 1; thence the Lemma concentrate itself on cases j > 1.
Now the prime numbers can be to build with generator form pn = 6n 1, therefore if there exist two
consecutive prime numbers in intervals I with k > 2 and j > 1 they have never D = 6j for the same
generator form.
However, if for absurd the consecutive prime numbers are such that:
pn+1

pn = 6j, j > 1 (2)

equivalent to:
pn

pn

6j

Then we conclude that

pn

pn

pn

6j

pn

(3)

Since the (2) is false, we cannot conclude the (3). In other words if the difference between two
consecutive prime numbers is multiple of 6 we have always found that the difference of square
roots of two consecutive prime numbers is greater than 1.
Andrica s conjecture
The Andrica s conjecture is true for consequence of Lemma 6 and Lemma 7.
Proof.
The conjecture supposes the existence of two consecutive prime numbers. If these are included in
the same square interval already the Lemma 5 will gives true the conjecture. If the two prime
6

numbers are included in different square intervals then Lemma 6 and Lemma 7 guarantee that the
conjecture is true.

2. The Cramer Shank Conjecture [2]

In this Section we have described the Cramer Shank Conjecture, utilizing the mathematics used in
the precedent Section on the Andrica s conjecture.
In the Cramer s conjecture, R(p) is the Cramer Shank ratio, it doesn t to be greater of 1 so that the
Cramer s conjecture is true, in other words Cramer s conjecture is true if
pn

R p

pn

ln p n

The greatest value of R(p) known is 0,92 for pn =1 693 182 318 746 370 with gap = 1132 between
this number and the following one pn+1 = pn + 1132.
It is interesting note that the value 0,92 is related to the aurea ratio by the following expression:

0,92

3
2

7/7

0,92705098

0,61803399

3
2

5 1
2

3
,
2

5 1
1,61803398...
2

where

Lemma
If the Andrica s conjecture is true, then Cramer s conjecture is true.
Proof.
In the precedent Section we have described some results concerning the Andrica s conjecture. Here
we have showed a consequence that influences on Cramer s conjectures
If the Andrica s conjecture is true, then is
pn

pn

From here, if we raise to square both the members and we take into consideration the algebraic
2
rule a b
a 2 2ab b 2 , we obtain
pn

pn
7

From here, we obtain:

pn

Re

pn

2 pn

pn

arranging the formula, subtracting to both the members pn we obtain

pn

pn

pn

pn 1 2 pn

R( p)

pn 2 pn

pn 1 pn
(ln pn ) 2

pn

2 pn

1 pn
pn

1 2 pn 2 pn
(ln pn ) 2

pn

For high value of prime number this is the demonstration that:

R( p)

pn 1 pn
(ln pn ) 2

3. On some equations concerning a proof of the Cramer s conjecture and the related
differences between prime numbers, principally the Cramer s conjecture and Selberg s
theorem. [3]

In number theory, Cramer s conjecture, formulated by the Swedish mathematician Harald Cramer
in 1936, states that
2
pn 1 pn O log pn , (3.1)
where pn denotes the nth prime number, O is big O notation, and log is the natural logarithm.
Cramer also gave much weaker conditional proof that
pn

pn

pn log pn

on the assumption of the Riemann hypothesis.

In the big-oh notation the eq. (3.1) can be rewritten also as follows
dn

Let us take f x

O log pn

log x . First we prove, that lim n

lim Tn

Sn
i

di
p1 log pi

(3.2)
Tn

S n is exists. We have that

di
.
log pi 1

(3.3)

We use the root test to show that the limit exists and is finite. The k th term is

ak

Let, vk

dk
pk

dk
log pk

dk
log pk

(3.4)

dk
, we get using the Bertrand s Postulate,
pk 1
p
pk
lim sup k 1
pk p k 1
k

1/ k

lim sup vk
k

2 1/ k

p
lim sup k
pk 1
k

Hence, looking at the conclusion of root test we can say, lim n

exists r0

N , such that for all n

1/ k

1.

Tn

(3.5)

S n exists. Therefore, there

r0

dx
x

dn
.
pn

Since, by the prime number theorem lim

x
x / ln x

pn

pn

(3.6)

1 , we can show that, lim n

pn 1 / pn

1.

Therefore, we get, as n
dn

pn 1
.
pn

pn 1 log

(3.7)

Therefore,
1/ k

lim sup ak

1/ k

dk
lim sup
log pk
k

dk
log pk

p
pk 1 log k 1
pk
lim sup
log pk
k

1/ k

p
pk 1 log k 1
pk
log pk 1

(3.8)

Now, we have, as x
x
log x

Also, we have p
theorem, lim n

1
n 1
n

1
pn
n

pn 1 / pn

. Also, as n

x .

, lim n

(3.9)

1
pn
n

1 . We know from the prime number

1 . Hence, as n
1

pnn

pn 1
pn

(3.10)

i.e,

n 1

log pn 1
log pn

n 2 .

Therefore, we get,
9

(3.11)

1/ k

lim sup ak

1/ k

pk 1
p
lim sup
log k 1
log pk 1
pk
k

1/ k

log pk 1
1
log pk

lim sup

pk

1/ k

p
lim sup k 1 log k 1
pk
k
This implies, lim n
have,

Tn

p
log k 1
pk

log pk 1
1
log pk

1/ k

log pk 1
1
log pk

p
lim sup log k 1
pk
k

1.

(3.12)

S n exists. Hence, we get, there exists n0 , such that for all n

dn
log pn

pn

dx
,
log x

pn

n0 , we

(3.13)

implies

dn
log pn

Li pn

Li pn . (3.14)

Similarly, we consider f x

log x . First we prove that lim n

lim Tn

Sn

i p1

di
log pi

di
log pi

Tn

S n exists. Here,

(3.15)

We use the comparison test to show that the limit exists. Here, the k th term is

dk
log pk

bk

dk
log pk

(3.16)

We can easily check that, as k

0 bk
(Since as n
Hence, the sum

, log pn
b

i 1 k

log pn

ak .

(3.17)

log pn log pn 1 ).

converges. This again implies, lim n

Hence, we get, there exists n1

N , such that for all n

dn
log pn

Li pn

pn
pn

dx
log x

Tn

S n exists, for f x

n1 , we have,

(3.18)

implies

dn
log pn

pn 1
log pn

Li pn

pn
,
log pn

where, Li x is the logarithmic integral function.

From equation (3.14) we get, there exists n2 N such that for all n
10

(3.19)

n2

log x .

dn
log pn

dn
log pn

Now, we have p
theorem, lim n

1
n 1
n

pn 1
log pn

1
pn
n

pn
log pn

pn 1
log pn

pn 1
log pn

, lim n

. Also, as n

pn 1
log pn

1
pn
n

log pn 1
1 .
log pn

(3.20)

1 . We know from the prime number

1 . Hence, as n

pn 1 / pn

1
n 1
n

pn 1
pn

(3.21)

i.e,

n 1
Hence, as n

log pn 1
log pn

n 2 .

(3.22)

from equation (3.20) we get

dn
log pn

pn

log pn 1
1
log pn

n 2

1.

(3.23)

n 2

n 1

n 1

Thence, we obtain the following expression:

dn
log pn
pn 1
log pn

log pn 1
1
log pn

dn
log pn

pn
2

pn

pn

dx
log x

log pn 1
1
log pn

1;

(3.24)

that can be rewritten also as follows:

dn
log pn
pn 1
log pn

log pn 1
1
log pn

dn
log pn

pn
2

pn

pn

dx
log x

log pn 1
1
log pn

5 1
1,61803398... (i.e., the aurea ratio), and
2
aurea section).
where

n 2

n 1

5 1
2

(3.25)

0,61803398... (i.e., the

3.1 The Cramer s conjecture and Selberg s theorem

Now we would like to consider a conjecture, due to H. Cramer, which is almost certainly true. The
conjecture is

11

lim sup
n

pn 1 pn
2
log n

1 . (3.26)

Obviously the conjecture implies that if k 1 and x x0 k , then the

contains a prime number.
The following theorem and its corollaries, provide some mathematical
Cramer s conjecture. For they imply that if the Riemann s hypothesis is
2
primes for which Pn 1 Pn is larger than log n is small . Let us
notation
h

d ,

Nh X

n
X Pn 2 X
dn h

1.

interval

x, x k log 2 x

support for a believe in

true, then the number of
introduce the following

(3.27)

Pn X
dn h

We can now state the principal result:

THEOREM 1
If the Riemann hypothesis is true, then

X
log 2 X ,
h

Nh X

X
log 2 X . (3.28)
2
h

COROLLARY
If the Riemann hypothesis is true, then
(i) d n

O Pn log Pn ,

d n2

(ii)

O X log3 X ,

(iii) If

4 , then

X Pn 2 X

d n2
log Pn
1
Pn

The above theorem is an elementary consequence of the following result.

THEOREM 2
Suppose that the Riemann hypothesis is true. If
0
X , then as X tends to infinity
X

x
x

1 dx

is a function of X such that

0 and

log 2 X

(3.29)

Deduction of theorem 1 from theorem 2. Let

0 be a fixed real number to be chosen later. We
1
shall consider two cases: (i) 0 h X
and (ii) X 1
h X . In case (i), we choose
h / 4 X and so 0
X . Now suppose that
p n , pn
and that d n

X ,2 X

h . If x satisfies
12

(3.30)

pn

1
d n , (3.31)
2

pn

then
x

1
dn
2

pn

pn

dn

pn

0.

(3.33)

(3.32)

with the consequence

x

1
dn
2

Hence, we have

1
dn
2

pn

1 dx .

pn

(3.34)

From theorem 2, we conclude that

dn
X pn 2 X
dn h

pn

2
X pn 2 X
dn h

1
dn
2

and choose h
h

log 2 X

1 dx

with 1 / 2

, then

X
log 2 X .
h

X 1 , we have proved that

X

x
x

pn

Thus if 0

2X

x
x

1 dx

(3.35)

X
log 2 X . However, if we take
1/ 2
h
X 0 for X X 0 . For if h X 0 , then

O
h

h , and since we are still in case (i), we also have

X
log 2 X ,
h

(3.36)

which leads to a contradiction if X is sufficiently large. Hence

for all h satisfying 0

X
log 2 X . From the definition of N h X , it
h

h Nh X

Nh

1
X
2

(3.37)

Upon replacing X by X / 2r , r 1,2,... and adding we deduce that

Nh X

X . Thus

is trivial that
X

X we have

0 for X

1
h

13

X
log 2 X .
2
h

(3.38)

Proof of the corollaries.

(i)

If we take h

(ii)

Nh X 0 .
We have

c X log X with c sufficiently large, it follows that N h X

dn

dn

1 h X X pn 2 X
dn h

d n2

X pn 2 X h d n

1 and so

(3.39)

X pn 2 X

and from theorem 1, we also have

dn

1 h X X pn 2 X
dn h

(iii)

X
log 2 X
h
X

O X log 3 X .

(3.40)

From (i), it follows that

X pn 2 X

d n2
log pn
pn

1
X log X

A
log X

d n2
X pn 2 X

(3.41)

Upon replacing X by 2r X for r 1,2,... and adding, we obtain

pn X

d n2
log pn
pn

r 1

log 2r X

r 1

r3

(3.42)

4.
and this latter series is convergent if
With regard the proof of theorem 3, we starts from the well known formula:

1
2 i

x
where Z S

log p p

Z S S
x dS ,
S

(3.43)

, and c denotes the line c it , c 1 . Now, being completely

formal, we move the line of integration to 1 / 2 z it , where z will be chose later, and encounter
a pole at S 1 with residue x . Taking a difference, we have
x

1
2 i

1
z
2

Z S
1
S

1 x S dS , (3.44)

thus
x

x
x

1
z
2

1
2

1
z it
2
1
it
2

14

1
z it
2

1 x it dt .

(3.45)

We now observe that the left hand side of the above equation is the formal Fourier transform of the
right hand side. From the Parseval inequality, we have
2
x

dx
x

1
z
2

1
2

1
2

z it

1
z it
2

1
2

1 dt .

(3.46)

z it

In fact, the above inequality does hold, but the rigorous argument, which closely parallels the above
formal manipulations, starts not with x but with a more artificial function which approximates to
x . However, assuming that the inequality has been proved, we see
2
x

dx
x

1
z
2

dx
x2 z

X 2z

x
x

1 dx ,

(3.47)
and so

X
0

1 dx

x2 z
2 2

1
2

z it

1
z it
2

1
2

1 dt . (3.48)

z it

Now we consider the integral on the right hand side of the above inequality. First of all, we note
that
S

Su S 1du

(3.49)

and
S

1 3,

(3.50)

since
1 and
1 . Thus, upon splitting the range of integration
,
to the three parts
, T , T, T , T,
and using the first estimate in the middle range and the second estimate
in the end ranges we obtain as an upper bound for the integral:

1
2

Z
T

1
2

z it
dt
z it

T
0

1
2

z it

It is now a relatively straightforward technical lemma to show that

15

dt .

(3.51)

1
2

Z
1
2

z it
dt

z it

1
, (3.52)
Tz 2

and
T
0

1
2

z it

dt

T
.
z2

(3.53)

X 2z
2
Tz 2

X 2 zT
, (3.54)
z2

Thus we now have

x

1 dx

and if we choose T 3 /
and z
completes the proof of theorem 2.

4 / log X , the upper bound becomes O log 2 X /

, which

4. On some equations concerning the p-adic strings and the zeta strings. [4] [5]
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:

A a, b

g2

a 1

1 x

g 2 DX exp

where

1 , T

1/

b 1

i
2

a b
a b

g2

dx

b c
b c

c a
c a

g2

1 a
a

1 b
b

1 c
c

d2

d2

exp ik j X

(4.1

4.4)

j 1

, and a

s
, b
2

t , c

with the condition

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

A a, b

g2

a 1

1 x

b 1

dx ,

is:
Ap a, b

g 2p

a 1

Qp

b 1

x p 1 x p dx ,

(4.5)

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
16

ax 2

bx d x
p P

Qp

ax 2

bx d p x 1 ,

a Q ,

b Q , (4.6)

what follows from

Qv

ax

bx d v x

a 2a v

1
2

b2
, v
4a

,2,..., p... . (4.7)

These Gauss integrals apply in evaluation of the Feynman path integrals

1
h

x ' ', t ' '

K v x' ' , t ' ' ; x' , t '

x ', t '

t ''
t'

L q, q, t dt Dv q ,

(4.8)

for kernels K v x' ' , t ' ' ; x' , t ' of the evolution operator in adelic quantum mechanics for quadratic
Lagrangians. In the case of Lagrangian
q2
4

1
2

L q, q

q 1 ,

for the de Sitter cosmological model one obtains

K x ' ' , T ; x ' ,0

x' ' , x' ,

1,

K p x ' ' , T ; x ' ,0

Q, T

Q ,

(4.9)

p P

where
2

T3
24

K v x' ' , T ; x' ,0

8T 4T v 2

x' ' x'

T
4

x' ' x'

8T

(4.10)

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:

K v x' ' , T ; x' ,0

8T 4T v

4 anti log

T3
24

cos txw'
e
cosh x
e

log

1
2

t2
w'
4
w'

10 11 2
4

x 2 w'

T
2
4

x' ' x'

dx

itw'

142
t 2 w'
.

(4.10b)

10 7 2
4

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

xp
p P

x ,x

0, x Q \ Z

17

(4.11)

x' ' x'

8T

xp

where

1 if x p

1 and

xp

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:
a

a 1

1 a
,
a

xd x

B a, b

a 1

1 x

b 1

a 1

B p a, b

Qp

a 1

xp

Qp

d x

b 1

x p 1 x p dpx

1 pa 1
,
1 p a

x dpx

c ,

(4.13)

c ,

(4.14)

(4.12)

where a, b, c C with condition a b c 1 and

a is the Riemann zeta function. With a
regularization of the product of p-adic gamma functions one has adelic products:
u

1 , B a, b

p P

1, u

B p a, b

0,1, u

a, b, c,

(4.15)

p P

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

1
1 p

exp

x x

a 1

d x
1
1 p
a

a 1

1 Q
p

x p x p dpx

a
2

a
,
2
a

(4.16)

, Re a 1 ,
a ,

(4.17)

(4.18)

p P

one obtains
A

where

1 a

a ,

(4.19)

a can be called adelic zeta function. We have also that

a

p P

Let us note that exp

x 2 and

x2 x

exp

a 1

d x

1
1 p

a 1

1 Q
p

x p x p dpx .

(4.19b)

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:
1
A

24 e

x2

xp
p P

whose the Fourier transform

18

(4.20)

kx

24 e

k2

kp
p P

has the same form as

x . The Mellin transform of

x x dAx

x x

a 1

1
d x
p
p P1

x is

xp x

1 Q
p

(4.21)

a 1

dpx

a
2

a
2

(4.22)

and the same for A k . Then according to the Tate formula one obtains (4.19).
The exact tree-level Lagrangian for effective scalar field
which describes open p-adic string
tachyon is

1 p2
g2 p 1

Lp

1
p
2
2
t

where p is any prime number,

p 1

p 1

(4.23)

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

Cn Ln

L
n 1

n 1
Ln
2
1 n

1
g2

1
2

n 1

n 1

1 n 1

(4.24)

Recall that the Riemann zeta function is defined as

s
n

1
s
1 n

1
1 p

, s

i ,

1.

(4.25)

Employing usual expansion for the logarithmic function and definition (4.25) we can rewrite (4.24)
in the form
1 1
L
ln 1
, (4.26)
g2 2
2
where

1.

where

acts as pseudodifferential operator in the following way:

ikx

1
2

eixk

k2 ~
k dk ,
2

x dx is the Fourier transform of

k2

k02

k2

(4.27)

x .

Dynamics of this field 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 is an open scalar zeta string. The equation of
motion for the zeta string is
19

1
2

k2 ~
k dk
2

eixk

k 02 k 2 2

(4.28)

0.
which has an evident solution
For the case of time dependent spatially homogeneous solutions, we have the following equation of
motion
2
1
k02 ~
t
t
t
e ik 0 t
k0 dk0
. (4.29)
k0
2
2
2
2
1
t
With regard the open and closed scalar zeta strings, the equations of motion are
1
2

1
4

k2 ~
k dk
4

eixk

n n 1
2

k2 ~
k dk
2

ixk

(4.30)

n 1

nn 1
2n 1

n2
n 1

n n 1
1
2

1 ,

(4.31)

0.

and one can easily see trivial solution

5. On some equations concerning the

gauge theory.

-deformation in toroidal compactification for N = 2

We denote the torus as To and endow it with a constant metric GIJ , I , J

dsT2o

n 1

Gij d i d

1,2 :

2 . (5.1)

The gauge theory probes the dual torus ToV , the moduli space of flat U 1 -connections on To . We
write such a connection as
A i 1d 1 i 2 d 2 , (5.2)
with constant hermitian matrices 1 ,
functions
un1 , n2
shift the components

. The gauge transformations generated by the U 1 -valued

exp in1

in2

(5.3)

by n1, 2 , respectively. The natural metric on ToV is given by:

1, 2

dsT2V
o

1
2 i

To

dA

det G G ij d i d

dA

(5.4)

It depends only on the complex structure of To . It is convenient to parametrize G ij by two complex

numbers

1, 2

,
20

G ij d i d

(5.5)

defined up to a simultaneous phase rotation, so that the invariants are:

2

G11 ,

G 22 ,

Re

G12 .

(5.6)

Let us assume
2

Im

We then have:

det G

Im

.
2

(5.7)

For a rectangular torus,

1
1

Let us first consider the case of gauge group U 1 . We take the Maxwell action to be
1
8

X4

4
Fmn F mn
2
g4

d 4x g

i
4

mnpq

F mn F pq .

(5.8)

If we take the four-manifold to be

To , with the product metric h G , with h being the metric
on , and denote the Riemannian measure of as d , then, in the low-energy approximation,
(5.8) reads as:
I

2
8

det g

8 ab ij
h G
g 42

det G
g 42

Ai

Aj

1 1

ab ij
a

Ai

Aj

1 1

(5.9)

We note that
2
8

39,4784176
1,57096 .
25,1327412

Now, we have that:

14 / 7

1,52786405

7/7

1
2

28 / 7

42 / 7

4
9

1,52786405 1,61803399

21

4
1,52786405 ;
9
1
3,14589804
1,57294902 .
2

3,437694
1
2

5 1
1,61803399 is the aurea ratio.
2

where

Thence, we can rewrite the eq. (5.9) also as follows:

14 / 7

det g

7/7

8 ab ij
h G
g 42

The bosonic part of the pure N

low energies by
L

8 2
det G tr
2 g 42

28 / 7

Ai

det G
g 42

4
9

42 / 7

Aj

ab

1
2

ij
a

1 1

Ai

Aj

1 1

(5.9b)

2 gauge theory Lagrangian reduced on the torus To is given at

i trd

(5.10)

where tr denotes the induced metric on t .

We note that
8 2
39,478 39,624
2
35 / 7

35 / 7

where

3
3
12,708204
19,06230590 ;
2
2
3
3
14 / 7
13,708204
20,56230590 ; 19,06230590 + 20,56230590 = 39,6246118;
2
2
5 1
1,61803399 is the aurea ratio.
2

Thence, we can rewrite the eq. (5.10) also as follows:

35 / 7

3
2

35 / 7

3 1
det G tr
2 g 42

14 / 7

i trd

(5.10b)

The gauge theory part of this Lagrangian can be borrowed from (5.9). We view here
t R / cwt Z as real, and
t C as complex, with tr defining a positive definite
1, 2
inner product on t . The Euclidean path integral measure is given by
e

(5.11)

22

The condition for a field configuration to be invariant under the supercharge Q that is relevant to
Donaldson theory and the -deformation is

0,

1 1

0,

(5.12)

where the second equation is anti-selfduality of the gauge field in our low energy approximation.
For such fields, (5.11) evaluates to:

exp

exp 2 i

trd

(5.13)

where the complexified gauge coupling is equal to

4 i
. (5.14)
g 42

For the eq. (5.10), thence, we can rewrite the eq. (5.13) also as follows:
8 2
det G tr ( 1d
2 g 42

exp

exp 2 i

( 1d

4 i
g 42

trd

) d

(5.14b)

(5.15)

i trd

Now we use the following notation:

8 2
det G
g 42

Im
Im

0
2

The Lagrangian (5.10) describes a sigma-model with target the product of a torus and t C , all
divided by the Weyl group. Upon T -duality along the 1 direction, we map it to a sigma-model on
M H , after taking into account the nonlinear corrections. The T -duality is performed in the standard
fashion. The first step is to replace d 1 in (5.10) by an independent t -valued one-form p1 and add
the term
2 itr p1 d ~1 to L , with the understanding that ~1 takes values in a circle of
circumference 1:
L'

tr

1 1

1 1

d ~1

2 itr

p1 .

(5.16)

Integrating over ~1 would lead us back to (5.10). Instead, we integrate over p1 . The path integral
over p1 is Gaussian, with the saddle point for p1 at:

p1

Re

2
1

2
2
0

d ~1

In terms of the left- and right-moving components of

23

(5.17)

, (5.17) reads as follows:

L
1

~
1

2
0

Re

R
1

~
1

Re

. (5.18)

The T-dual Lagrangian is given by:

tr

d ~1

Im

1
2

2
0

2 i Re

tr d ~2

d ~1

2
1

(5.19)

C -valued dimensionless coordinates Z ,W :

Introduce the t

4 i
g 42

~
1

1
2

(5.20)

In terms of W and Z , eq. (5.19) takes the form:

2

2
0

tr dZ

dZ

dW

dW

2 i Re

tr d ~1

(5.21)

Note that
Im

1
2

We note also that

24
log
142

Im

(5.22)

can be expressed also in the following form (Ramanujan modular equation):

10 11 2
4

10 7 2
4

. We observe that 24 is the number concerning the

modes of the physical vibrations of the bosonic strings. Furthermore, we observe also that the
Fibonacci zeta function is

f n s , where the n th Fibonacci number can be expressed as

s
n 1

fn

5
function is:

and where

'F s

5 / 2 is the aurea ratio. The derivative of the Fibonacci zeta

s2

1
2 ln
24

1
ln

3 ln 2 5
ln

24

6 ln 5

ln c

s .

Also here, we note that there is the number 24, i.e. the modes corresponding to the physical
vibrations of the bosonic strings.
Thence, from (5.21) and (5.22) we obtain the following mathematical connections with the
Ramanujan modular equation, the Fibonacci zeta function and the Palumbo-Nardelli model:

24
log
142

2
0

R
16 G

d 26 x g
1
2
10

2
0

tr dZ

1/ 2

dZ

dW

dW

2 i Re

24
log
142

1
ln

2
10

2
10
2
10

s2

R
16 G

d 10 x

1/ 2

6 ln 5

1
g g Tr G G
8

Im

(5.22b)

2
1

Im

ln c

1
g
2

1 ~ 2
H3
2

R 4

Tr F2

tr d ~1

10 7 2
4

3 ln 2 5
ln

1
2 ln
24

d 26 x g
1

10 11 2
4

1
g
2

1 ~ 2
H3
2

R 4

Im

1
g g Tr G G
8

ln c

d 10 x

6 ln 5

Im

10 7 2
4

3 ln 2 5
ln

1
2 ln
24

s2

ln

10 11 2
4

2
10
2
10

Tr F2

(5.22c)

We deduce from (5.21) the target space metric

dsM2 H

Im 2 /
Im 0

dZdZ

dWdW .

(5.23)

In our approximation, the target space metric is flat; in the exact theory, it is a complete hyperKahler metric on what we usually call M H . We also deduce from (5.21) a B-field, which, up to
exact terms, is given by:
25

2 Re 2 /
2i Im 0

dZ

dZ

dW

dW

Re

(5.24)

Here I is the topologically normalized symplectic form on M H , which is Kahler in the complex
structure I . The functions of Z ,W are holomorphic in complex structure I .
Also here we can note that there exists the mathematical connection with the Aurea section. Indeed,
we remember that that is present in many equations of this chapter, is related to the Aurea section
5 1
by the following simple but fundamental relation:
2

arccos

0,2879

6. Mathematical connections

Now we take the eq. (3.25). We obtain the following connections with the eqs. (4.19b) and (5.9):

dn
log pn
pn 1
log pn
A

log pn 1
1
log pn

dn
log pn
p

2
8

det g

det G
g 42

Ai

Aj

1 1

n 2

a 1

x2 x

exp

8 ab ij
h G
g 42
2

log pn 1
1
log pn

p P

dx
log x

pn

pn

pn
2

1
1 p

d x

ab

n 1

ij
a

a 1

x p x p dpx

1 Q
p

Ai

1 1

Aj

(6.1)

Now we take the eqs. (3.46), (3.48) and (3.54). We obtain the following connections with the eq.
(4.28):

1
z
2

dx
x

1
2

1
k 02 k 2 2

1
2
1
2

eixk

26

z it
1

1
z it
2

1 dt

z it

k2 ~
k dk
2

(6.2)

x2 z
2 2

1 dx

1
2

1
2

z it

1
z it
2

k2 ~
k dk
2

ixk

1 dx

eixk

k 02 k 2 2

1 dt

z it

1
2

k 02 k 2 2

1
2

X 2z
2
Tz 2

X 2 zT
z2

k2 ~
k dk
2

(6.3)

(6.4)

In conclusion, we take the relationship (5.22c) that can be connected with the eq. (4.28) as follows:

LT
2

2
0

tr dZ

dW

dW

2 i Re

tr d ~1

24
log
142

2
0

dZ

1
ln

10 11 2
4

1
2 ln
24

s2
1

k 02 k 2 2

10 7 2
4

3 ln 2 5
ln
eixk

Im

Im

6 ln 5

ln c

k2 ~
k dk
2

(6.5)

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 and Prof. Ignazio Licata of Institute
for Scientific Methodology (Palermo, Italy) for his useful discussions

27

References

[1] Turco Rosario, Colonnese Maria, Nardelli Michele, Di Maria Giovanni, Di Noto Francesco e
Tulumello Annarita (2010) Miglioramento e Nota correttiva Proposta di soluzione congettura di
Andrica. Solar CNR e sito www.gruppoeratostene.it.
[2] Turco Rosario, Colonnese Maria, Nardelli Michele, Di Maria Giovanni, Di Noto Francesco e
Tulumello Annarita (2010) Proposta di Dimostrazione Congettura di Cramer - Shank. Solar CNR
e sito www.gruppoeratostene.com.
[3] William J. Ellison Differences between prime numbers Sminaire Delange-Pisot-Poitou.
Thorie des nombres, tome 15, n 1 (1973-1974), exp. n 18, p. 1-10.
[4] Branko Dragovich: Adeles in Mathematical Physics
2007.
[5] Branko Dragovich: Zeta Strings

arXiv:0707.3876v1 [hep-th] 26 Jul

arXiv:hep-th/0703008v1

1 Mar 2007.

[6] Nikita Nekrasov and Edward Witten The Omega Deformation, Branes, Integrability, And
Liouville Theory arXiv:1002.0888v1 [hep-th] 4 Feb 2010.

28

This document was created with Win2PDF available at http://www.win2pdf.com.

The unregistered version of Win2PDF is for evaluation or non-commercial use only.