The Circle’s Method to investigate the Goldbach’s Conjecture and the Germain primes: Mathematical

connections with the p-adic strings and the zeta strings.

Michele Nardelli 1, 2 e Rosario Turco

1
Dipartimento di Scienze della Terra

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

80138 Napoli, Italy

2
Dipartimento di Matematica ed Applicazioni “R. Caccioppoli”

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 we have described in the Section 1 some equations and theorems concerning the
Circle Method applied to the Goldbach’s Conjecture. In the Section 2, we have described some
equations and theorems concerning the Circle Method to investigate Germain primes by the
Major arcs. In the Section 3, we have described some equations concerning the equivalence
between the Goldbach’s Conjecture and the Generalized Riemann Hypothesis. In the Section 4, we
have described some equations concerning the p-adic strings and the zeta strings. In conclusion, in
the Section 5, we have described some possible mathematical connections between the
arguments discussed in the various sections.

1. On some equations and theorems concerning the Circle Method applied to the Goldbach’s
Conjecture [1]

Generating Functions on the Circle Method

If we consider a generating function of a power series with z0=0 and ρ=1 we have that:

1
 ∞
a =1 se a n ∈ N
f ( z ) = ∑ an ⋅ z n  n
n =0 0 altrimenti

Often are interesting only the representations at k at a time of numbers ai, the sum of these
returns another number n; for which the k-ple obtained, as part of the set of all integers N, are
only a subset of the Cartesian product Nk (or of N x N if k=2).

For example, for k=2 if n is the number sum belonging to N (set of the integers), and a1, a2 the
numbers belonging to N, then the representations in pairs of numbers ai are:

r 2 (n) := {(a1 , a2 ) ∈ N × N : n = a1 + a2 } (1)

In general with the Taylor’s series or with the residue Theorem, we obtain that:

1 f ( z)
r2 (n) =
2π i ∫ z n +1
dz (2)

i.e. r2(n) correspond with an of the series.

Having to work with a Cartesian product, it is possible also define (for the Cauchy’s product):
∞ ∞
f 2 ( z ) = ∑ cn ⋅ z n cn = ∑ ∑ ah ⋅ ak z n (3)
n=0 h ≥0 k ≤n
h+ k =n

with akah≠1 if h and k belonging to N, then cn correspond to r2(n); thence, also here we have that:

1 f 2 ( z)
r2 (n) =
2π i ∫ z n +1
dz (4)

where the integral on the right is along a circle γ(ρ), path counterclockwise, centered in the origin
and unit radius.

Is logic that in the case of Goldbach’s Conjecture we replace at N the set of primes P and the terms
ai are belonging to P.

In general we have to deal with an additive problem with k>2, such as: the Waring’s problem (any
k>2 and power s); the Vinogradov’s Theorem (for k=3 and with ai belonging to the set of the prime
numbers P); the problem of the twin number for n=2, if –P={-2,-3,-5,-7,-11,…} and we consider P-
P, studying the pairs (p1,p2) with p1,p2∫P such that n=p1-p2.

Then in the general case for k>2 is interesting the resolution of the equation of the type:

n = a1 + a2 + ... + ak

r k (n) := {(a1 , a2 ,..., ak ) ∈ N k : n = a1 + a2 + ... + ak }

In this case we must choose a function

2
 ∞
f ( z ) = ∑ rk (n) z n
s

n=0

for which:

1 f s ( z )dz
rk (n) =
2π i ∫ z n +1
(4’)

If we choose as generating function of the series the

∞
1
f ( z) = ∑ z n =
n =0 (1 − z ) −1

Thence, if the power of the f is s=1, the eq. (4’) can be rewritten as follow:

1 dz
rk (n) =
2π i ∫ (1 − z ) k
z n +1
(5)

The eq. (5) of the general case is interesting because the integrand function has a easy singularity
on γ(1); so it can easily integrate. In fact, for ρ<1 we can develop the k-th power of a binomial:

1  −k   −k   −k   −k  ∞
 −k 
= (1 − z ) − k =   (− z )0 +   (− z )1 +   (− z )2 + ... +   (− z )m = ∑  (− z ) m
(1 − z )
k
0  1  2  m  m =0  m 

If we replace in the eq. (5), we obtain:

1 dz 1 ∞  −k 
∫ (1 − z )k z n+1 2π i m∑=0  m (−1)
m − n −1
rk (n) = = ∫z dz =
m

2π i

For m = n the integral is 2πi, while 0 in the other cases; thence the result is:

 − k   n + k − 1
= (−1)n   =   (6)
 n   k −1 

Thence we can rewrite the eq. (6) as follows:

1 dz 1 ∞  −k 
∫ (1 − z )k z n+1 2π i m∑=0  m (−1)
m − n −1
rk (n) = = ∫z dz =
m

2π i

 − k   n + k − 1
= (−1)n   =   (6b )
 n   k −1 

The eq. (6) represent just the way of writing n as a sum of k elements of power s=1.

3
The Vinogradov’s simplification

Vinogradov, however, made the further observation that to r2(n) can contribute only the integers
m≤n; so we can introduce a different function (as opposed to one that led to eq. (6)), more useful
to:

N
1 − z N +1
f N ( z ) := ∑ z m = true for z≠1
m=0 1− z

For n≤N, for the Cauchy’s Theorem, we can write:

1 f N ( z )k dz
rk (n) =
2π i ∫ z n +1
(7)

Now fN(z) is a finite sum and there are no convergence problems, so the integrand function in the
equation (7) has no points of singularity. Thence, now we can permanently fix the curve on which
it integrates. Indeed, we take as curve the complex exponential function e( x) := e2π ix (1) and making
a change of variable z = e(α) the eq. (7) becomes:

∫ V (α ) e(−nα )dα = ∑ ∑ e(( p + p − n)α )dα = r2 (n)

2
1 2 (8)
0 p1 ≤ N p2 ≤ N

The eq. (8) is the n-th Fourier coefficient of f Nk (e(α )) . If now we put T (α ) = f N (e(α )) , we obtain:

 1
N 1 − e(( N + 1)α ) e( 2 Nα ) sin(π ( N + 1)α )
T (α ) = T N (α ) = f N (e(α )) = ∑ e(mα ) =  = αε Z
1 − e (α ) sin( πα )
m =0 
 N + 1 αε Z

The function above now can be studied and is an element of elemental analysis exploitable.

Property of TN(α)

The function T has peaks on the integer values and decreases much on the non-integer values. It’s
easy to see, also as numerical computation, (see [3]) that:

1 1
| T N (α ) |≤ min( N + 1, ) ≤ min( N + 1, ) (9)
| sin(πα ) | || α ||

0
1 se a = b
1
We note that this function is orthogonal in the range [0,1] indeed: ∫ exp(az ) exp(−bz )dx = 0 altrimenti
1
for

0
which is rk (n) = ∫ f k (exp( z )) exp(− z )dz
1
4
where ||α|| is the distance between two numbers, T is periodic of period 1 and α<sin(πα) for
αε(0,1/2]. If we use the eq. (9) and δ = δ(N) is chosen not too small, then the contribution in the
range [δ,1-δ] is negligible. For example, if δ>1/N then:
1−δ 1−δ 1−δ
dα 2 1− k
| ∫δ TNk (α )e(− nα )dα |≤ ∫ | TNk (α ) | dα ≤ ∫ ≤ δ (10)
δ || α || k −1
k
δ

Taking into account of eq. (8) and eq. (10) for n = N, k=2 e δ −1 = o( N ) we obtain that:

1−δ 2 2
1  sin(π ( N + 1)α )   sin(π ( N + 1)α ) 
1
r2 (n) = ∫   d α = 2 ∫   dα (10b)
2π i 0  sin(πα )  δ  sin(πα ) 

practically the Goldbach’s Conjecture via the Circle Method.

The Goldbach’s Conjecture, without to consider the difference between the weak and the strong
conjecture, says that “given an even number greater than 4 this is always the sum of two prime
numbers”.

For the Goldbach’s Conjecture, therefore, we are interested to the representations:

r 2 (n) := {( p1 , p2 ) ∈ P × P : n = p1 + p2 }

where p1, p2 are prime numbers not necessarily distinct, belonging to the set of the prime
numbers P and for the moment we do not consider n as even number, but any (for example we
accept also 2+3=5 at this stage of investigation). Putting:

V (α ) = VN (α ) = ∑ e( pα )
p≤ N
(11)

then the Goldbach’s problem, with the techniques of real and complex analysis, results for n ≤ N:
1 1

∫ V (α ) e(−nα )dα = ∑ ∑ ∫ e(( p + p2 − n)α )dα = r2 (n) (12)

2
1
0 p1 ≤ N p2 ≤ N 0

In the following, instead of consider directly the eq. (12), we can consider a weighted version with
weight different from 1 (instead of consider p1+p2, we consider log(p1+p2)=logp1*logp2):

R2 (n) := ∑
p1 + p2 = n
log p1 log p2 (13)

It’s clear that r2(n) is positive if R2(n) is also positive; then it is sufficient to study R2(n) for the
Goldbach’s conjecture.

A weighted version of eq. (11) now is:

5
 S (α ) = S N (α ) = ∑ log p e( pα )
p≤N
(14)

Bearing in mind the Dirichlet’s Theorem on the arithmetic progressions, choosing q, a such that
MCD(q,a)=1, we write that

θ ( N ; q, a ) = ∑
p≤ N
log p (15)
p ≡ a mod q

Theorem of Siegel - Walfisz

Let C,A>0 with q and a relatively prime, then

N N
∑
p≤ N
log p =
ϕ (q)
+ O( C )
log N
p ≡ a mod q

for q ≤ log A N

the previous constant C does not depend on N, a, q (but more depend on A: C(A)).

Thence, from the Theorem of Siegel – Walfisz (see [4]) we have that:

N N
θ ( N ; q, a ) = + O( C ) (16)
ϕ (q ) log N

where we have defined

N
E ( N ; q, a ) = O( ) = O( N exp(−C ( A) log N )) .
log C N

where ϕ is the Euler totient function and C must be chosen not very large. The theorem is
effective when q is very small compared to N. At this point, similarly to (12) we can write that for
n≤N:
1

∫ S (α ) e(−nα )dα
2
(17)
0

As preliminary operation, we see some value of S (for the exponential we recall the transformation
x → exp(2π ix) thence for example to ½ it comes to exp(2πi*1/2)=-1):

S(0) = θ(N,1,1)≈N

S(1/2) = - θ(N,1,1)+2 log2≈-N

S(1/3) = exp(1/3) θ(N,3,1) + exp(2/3) θ(N,3,2) + log 3 ≈ - ½ N

6
S(1/4) = exp(1/4) θ(N,4,1) + exp(3/4) θ(N,4,3) + log 2 ≈ 0

Now we see S also for some rational value a/q, when 0 ≤a≤q and MCD(a,q)=1. In this case the
eq. (14) becomes:
q
S (a / q) = ∑ ∑ log p e( p ⋅ a / q ) =
h =1 p ≤ N
p ≡ h mod q

q q q
= ∑ e(h ⋅ a / q ) ∑ log p = ∑ e(h ⋅ a / q ) ⋅ θ ( N ; q, a ) = ∑ * e(h ⋅ a / q ) ⋅ θ ( N ; q, a ) + O (log q log N ) (18)
h =1 p≤ N h =1 h =1
p ≡ h mod q

the asterisk in the last summation denote the further condition that MCD(h,q)=1.

From the eq. (18) taking into account the eq. (16), we obtain:

N q * q
S (a / q) = ∑
ϕ (q) h =1
e ( h ⋅ a / q ) + ∑
h =1
*
e(h ⋅ a / q ) ⋅ E ( N ; q, a ) +O (log q log N )

µ (q) q *
= + ∑ e(h ⋅ a / q ) ⋅ E ( N ; q, a ) +O (log q log N ) (19)
ϕ (q) h =1

where µ is the Moebius’s function(2).

For the Moebius’s function, |S(α)| is large when α is a rational number, in a neighborhood of a/q,
and from the previous examples we have seen also that S(a/q) decreases as 1/q.

Realizing how about S, we can now try to find an expression for R2(n) and usually is used the
“partial sum on the arcs”.

a
Putting: α = + η , for |η| small, we obtain:
q

µ (q) µ (q)
S (a / q + η ) = ∑
ϕ (q ) m≤ N
e(mη ) ⋅ E ( N ; q, a,η ) =
ϕ (q)
T (η ) + E ( N ; q, a,η )

From the eqs. (16) and (19) we have that:

E ( N ; q, a,η ) = OA (q(1 + N | η |) N exp(−C ( A) log N ))

2
µ(q)=0 se q è divisibile per il quadrato di qualche numero primo, è (-1)k se q=p1p2...pk dove i pi sono k numeri primi
distinti.
7
 a a
If as in [3] we denote with M (q,a) := ( − ξ ( q, a ), + ξ '(q, a )) the Farey’s arc concerning the rational
q q
number a/q, with ξ ( q, a ) and ξ '( q, a ) of order (qQ)-1, then we define the set or the union of the Major
and Minor arcs as follow:

q
M := U U *M (q, a ) m:= [ξ (1,1),1 + ξ (1,1)] \ M (20)
q ≤ P a =1

Also here the asterisk indicates the additional condition that the MCD(q,a)=1. For the range of the
Minor arc instead of consider [0,1] we have passed to [ξ (1,1),1 + ξ (1,1)] , that is possible for the
periodicity 1.

It’s clear that, starting again to the eq. (17), now R2(n) is the sum of two integrals, one on the
Major arc and the other on the Minor arc (as we have said when we have considered the eq. (12))
and for n≤N is:
1
R2 (n) = ∫ S (α )2 e(− nα )dα = ( ∫ + ∫ )S (α )2 e(− nα )dα =
0 M m

How we have defined the Major arcs in the eq. (20), we have that

q ξ '( q , a )
a a
R2 (n) = ∑∑ *
∫ S ( + η ) 2 e(− n( + η ))dη + ∫ S (α ) 2 e(− nα )dα = RM (n) + Rm (n) (21)
q ≤ P a =1 − ξ( q,a )
q q m

In the following with the symbol ≈ we denote as in [3] an asymptotic equality (to the infinity).

The eq. (21) can be rewrite also as follows:

ξ '( q , a )
q
µ (q) 2 a
RM (n) ≈ ∑∑ * ∫ T (η ) 2 e(− n( + η ))dη =
q ≤ P a =1 −ξ ( q , a )
ϕ (q) 2
q

ξ '( q , a )
µ (q)2 q * a
=∑ 2 ∑
e(− n ) ∫ T (η ) 2 e(− nη )dη (22)
q ≤ P ϕ ( q ) a =1 q −ξ ( q , a )

If we extend the integral that contains T throughout the range [0,1]

1

∫ T (η ) e(−nη )dη = ∑ 1 = n − 1 ≈ n (23)

2

0 m1+ m 2 = n

Thence, we obtain:

µ (q)2 q * a
RM (n) ≈ n ∑ 2 ∑
e( − n ) (24)
q ≤ P ϕ ( q ) a =1 q

8
where the inner sum is called the Ramanujan’s sum and we can show it with a Theorem that can
be expressed as a function of µ and ϕ:

q
µ(
)
µ (q)2
q ϕ (q) µ (q) 2
( q, n)
RM (n) ≈ n ∑ µ( ) = n∑
q ≤ P ϕ (q) q ≤ P ϕ (q)
2
( q, n) ϕ ( q ) ϕ(
q
)
( q, n) ( q, n)

If we extend the sum to q≥1 and we consider another Theorem (see [3]), we obtain:

q
µ( )
µ (q) 2
( q, n)
R2 (n) ≈ n ∑ = n∏ (1 + f n ( p )) (25)
q ≥1 ϕ ( q )
q
ϕ( ) p
( q, n)

The “productor” is on all the prime numbers; further we have that:

q  1
µ( )  se p|n
µ (q )  p −1
2
( q, n)
f n ( p) = =
ϕ (q) ϕ ( q )  1
− altrimenti
( q, n)  ( p − 1) 2

If n is odd then 1 + f n (2) = 0 thence the eq. (23) states that there aren’t Goldbach’s pairs for n.

Indeed R2(n)=0 if n-2 is not prime number, R2(n)=2log(n-2) if n-2 is a prime number. If, instead, n is
even, we can obtain the following expression:

1 1
R2 (n) ≈ n∏ (1 + )∏ (1 − )
p −1 p |n ( p − 1)
2
p |n

p ( p − 1)
2
1 p −1
R2 (n) ≈ 2n∏ ( ⋅ )∏ (1 − ) = Co n∏ ( ) (26)
p − 1 p( p − 2) p > 2 ( p − 1) p−2
2
p| n p |n
p>2 p >2

where C0 is the constant of the twin primes. The eq. (26) is the asymptotic formula for R2(n) based
on the Number Theory and provides a value greater than r2(n) of a quantity ( log n )2, for the
weights logp1logp2.

2. On some equations concerning the Circle Method to investigate Germain Primes [2]

In this section we apply the Circle Method to investigate Germain primes. As current techniques
are unable to adequately bound the Minor arc contributions, we concentrate on the Major arcs,

9
where we perform the calculations in great detail. The methods of this section immediately
generalize to other standard problems, such as investigating twin primes or prime tuples.

We remember the Siegel-Walfisz Theorem, that will be useful in the follow.

Let C , B > 0 and let a and q be relatively prime. Then

x  x 
∑ log p =
φ (q )
+ O C  . (27)
p≤ x  log x 
p ≡ a (q )

Definition 1

p −1 p −1
A prime p is a Germain prime (or p and are a Germain prime pair) if both p and are
2 2
prime. An alternate definition is to have p and 2 p + 1 both prime.

Let B, D be positive integers with D > 2 B . Set Q = log D N . Define the Major arc Ma, q for each
pair (a, q ) with a and q relatively prime and 1 ≤ q ≤ log B by

  1 1 a Q
Ma , q =  x ∈  − ,  : x − <  (28)
  2 2 q N

a 1
if ≠ and
q 2

 1 1 Q   1 Q 1
M1, 2 = − ,− +  U  − ,  . (29)
 2 2 N   2 N 2

We have that the our generating function is periodic with period 1, and we can work on either
 1 1
[0,1] or − ,  . As the Major arcs depend on N and D , we should write Ma , q ( N , D ) and
 2 2
M ( N , D ) . Note we are giving ourselves a little extra flexibility by having q ≤ log B N and each
log D N
Ma , q of size . By definition, the Minor arcs m are whatever is not in the Major arcs. Thus
N
 1 1
the Major arcs are the subset of − ,  near rationals with small denominators, and the Minor
 2 2
arcs are what is left. Here near and small are relative to N . Then
1
r (1; A1N , A2 N ) = ∫ FN ( x )e(− x )dx = ∫ FN ( x )e(− x )dx + ∫ FN ( x )e(− x )dx . (30)
2
1
− M m
2

10
We chose the above definition for the Major arcs because our main tool for evaluating FN ( x ) is
the Siegel-Walfisz formula (see eq. (27)), which states that given any B, C > 0 , if q ≤ log B N and
(r , q ) = 1 then
N  N 
∑ log p =
φ (q )
+ O C  . (31)
p≤ N  log N 
p ≡ r (q )

For C very large, the error term leads to small, manageable errors on the Major arcs.

Now we apply partial summation multiple times to show u is a good approximation to FN on the
Major arcs Ma , q . Define

cq (a )cq (− 2a )
Cq (a ) = . (32)
φ (q )2

We show

Theorem 1

For α ∈ Ma, q ,

 a  N2 
FN (α ) = Cq (a )u α −  + O C − 2 D  . (33)
 q  log N

The problem is to estimate the difference

 a  a
S a , q (α ) = FN (α ) − Cq (a )u  α −  = FN  β +  − Cq (a )u (β ) . (34)
 q  q

N2
To prove Theorem 1 we must show that Sa , q (α ) ≤ . It is easier to apply partial
logC − 2 D N
summation if we use the λ -formulation of the generating function FN because now both FN and
u will be sums over m1 , m2 ≤ N . Thus

Sa , q (α ) = ∑ λ (m )λ (m )e((m 1 2 1 − 2m2 )β ) − Cq (a ) ∑ e((m − 2m )β )

1 2
m1 , m 2 ≤ N m1 , m 2 ≤ N

  a 
= ∑ λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a )e((m1 − 2m2 )β )
m1 , m 2 ≤ N   q 

   a  
= ∑  ∑ λ (m )λ (m )e (m 1 2 1 − 2m2 )  − Cq (a ) e(− 2m2 β )e(m1β )
m1 ≤ N  m2 ≤ N   q  
11
  
= ∑  ∑ a (m , N )b (m , N )e(m β )
m2 1 m2 1 1
m1 ≤ N  m2 ≤ N 

= ∑ S (α ; m )e(m β ) ,
m1 ≤ N
a,q 1 1 (35)

where

 a
am2 (m1 , N ) = λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a ) ; bm2 (m1 , N ) = e(− 2m2 β )
 q

Sa , q (α ; m1 ) = ∑ a (m , N )b (m , N ) .
m2 1 m2 1 (36)
m2 ≤ N

Recall the integral version of partial summation states

N

∑ a b(m ) = A(N )b(N ) − ∫ A(u )b' (u )du ,

N
m (37)
1
m =1

where b is a differentiable function and A(u ) = ∑ m ≤ u am . We apply this to am2 (m1 , N ) and
bm2 (m1 , N ) . As bm2 = b(m2 ) = e(− 2β m2 ) = e −4πiβm2 , b' (m2 ) = −4πiβ e(− 2β m2 ) . Applying the integral
version of partial summation to the m2 -sum gives

  a 
Sa , q (α ; m1 ) = ∑ λ (m )λ (m )e (m
1 2 1 − 2m2 )  − Cq (a ) e(− 2m2 β ) = ∑ am2 (m1 , N )bm2 (m1 , N ) =
m2 ≤ N   q  m2 ≤ N

  N  
=  ∑ am2 (m1 , N ) e(− 2 Nβ ) + 4πiβ ∫  ∑ am2 (m1 , N )e(− uβ )du . (38)
u =1
m2 ≤ N  m2 ≤u 

The first term is called the boundary term, the second the integral term. We substitute these into
(35) and find

  
S a , q (α ) = ∑   ∑ a (m , N )e(− 2 Nβ )e(m β ) +
m2 1 1
m1 ≤ N  m2 ≤ N  

   
∑ 4πiβ ∫ ∑ am2 (m1 , N ) e(− uβ )du  e(m1β ) . (39)
N
+ 
u =1
m1 ≤ N  m2 ≤ u  

The proof of Theorem 1 is completed by showing S a , q (α ; B ) and S a , q (α ; I ) , where B = Boundary

and I = Integral, are small. The first deal with the boundary term from the first partial summation
on m2 , S a , q (α ; B ) .

Lemma 1

12
  N2 
Sa , q (α ; B ) = O C − D  . (40)
 log N

Proof. Recall that

    
S a , q (α ; B ) = ∑   ∑ a (m , N )e(− 2 Nβ )e(m β ) = e(− 2 Nβ ) ∑  ∑ a (m , N )e(m β ) .
m2 1 1 m2 1 1 (41)
m1 ≤ N  m2 ≤ N   m1 ≤ N  m2 ≤ N 

As e(− 2 Nβ ) = 1 , we can ignore it in the bounds below. We again apply the integral version of
partial summation with

  a 
am1 = ∑ a (m , N ) = ∑ λ (m )λ (m )e (m
m2 1 1 2 1 − 2m2 )  − Cq (a ) ; bm1 = e(m1β ) . (42)
m2 ≤ N m2 ≤ N   q 

We find

   
e(2 Nβ )S a , q (α ; B ) = ∑  ∑ a (m , N )e(Nβ ) − 2πiβ ∫ ∑  ∑ a (m , N )e(tβ )dt .
N
m2 1 m2 1 (43)
t =0
m1 ≤ N  m2 ≤ N  
m1 ≤ t m 2 ≤ N 

To prove Lemma 1, it suffices to bound the two terms in (43), which we do in Lemmas 2 and 3.

Lemma 2

   N2 
∑  ∑ m2 1  a (m , N ) e ( N β ) = O  C  . (44)
m1 ≤ N  m 2 ≤ N   log N 

Proof. As e(Nβ ) = 1 , this factor is harmless, and the m1 , m2 -sums are bounded by the Siegel-
Walfisz Theorem.

  a 
∑ ∑ a (m , N ) = ∑ ∑ λ (m )λ (m )e (m
m2 1 1 2 1 − 2m2 )  − Cq (a ) =
m1 ≤ N m 2 ≤ N m1 ≤ N m 2 ≤ N   q 

  a    a 
=  ∑ λ (m1 )e m1   ∑ λ (m2 )e − m2  − Cq (a )N 2 =
m1 ≤ N  q   m 2 ≤ N  q 

 c (a )N  N   c (− 2a )N  N 
= q + O C  ⋅  q + O C  − Cq (a )N 2 =
 φ (q )  log N   φ (q )  log N 

 N2 
= O C  (45)
 log N 

cq (a )cq (− 2a )
as Cq (a ) = and cq (b ) ≤ φ (q ) .
φ (q )2

13
Lemma 3

   N2 
( ) ( )
N
2πiβ ∫
t =0
∑  ∑ m2 1  a m , N e t β dt = O  C − D  . (46)
m1 ≤ t  m 2 ≤ N   log N

Q log D N c (a ) c (− 2a )
Proof. Note β ≤ = , and Cq (a ) = q 2 q . For t ≤ N , we trivially bound the
N N φ (q ) φ (q )2
m2 -sum by 2 N . Thus these t contribute at most
N
β∫
t =0
∑ 2 Ndt = β N
m1 ≤ t
2
≤ N log D N . (47)

An identical application of Siegel-Walfisz as in the proof of Lemma 2 yields for t ≥ N ,

 c (a )t  t   cq (− 2a )N  N 
∑ ∑ a (m , N ) =  φ (q ) + O log
m2 1
q
C
 ⋅ 
N   φ (q )
+ O C  − Cq (a )tN =
m1 ≤t m2 ≤ N    log N 

 tN 
= O C
 . (48)
 log N 

Therefore

 N 3β   N2 
( )
N
β∫
t= N
∑ ∑ m2 1 a m , N dt = O  
 log C N  = O  C − D  . (49)
m1 ≤ t m 2 ≤ N    log N

We note also that:

   N2 
2πiβ ∫ ∑  ∑ am 2 (m1 , N )e(tβ )dt = O C − D  ⇒
N

t =0
m1 ≤ t  m 2 ≤ N   log N

 N 3β   N2 
( )
N
⇒β∫
t= N
∑ ∑ m2 1 a m , N dt = O  
 logC N  = O  C − D  . (50)
m1 ≤ t m 2 ≤ N    log N

We now deal with the integral term from the first partial summation on m2 , S a , q (α ; I ) .

Lemma 4

 N2 
Sa , q (α ; I ) = O C − 2 D  . (51)
 log N

Proof. Recall

   
S a , q (α ; I ) = 4πiβ ∑ ∫ ∑ am 2 (m1 , N ) e(− uβ )du  e(m1β ) (52)
N

u =1

m1 ≤ N 
 m2 ≤u  

14
where

 a
am 2 (m1 , N ) = λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a ) . (53)
 q

Thence, the eq. (52) can be rewrite also as follow:

   a  
S a , q (α ; I ) = 4πiβ ∑ ∫ ∑ λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a )e(− uβ )du  e(m1β ) . (53b)
N

u =1

m1 ≤ N 
  m2 ≤ u  q  

We apply the integral version of partial summation, with

N  
am1 = ∫  ∑ am 2 (m1 , N ) e(− uβ )du bm1 = e(m1β ) . (54)
u =1
m2 ≤u 

We find

 
S a , q (α ; I ) = 4πiβ  ∑ ∫ ∑ am 2 (m1 , N )e(− uβ )du  e( Nβ ) +
N

u =1
 m1 ≤ N m2 ≤ u 

N  
+ 8πβ 2 ∫  ∑ ∫ ∑ am2 (m1 , N )e(− uβ )du e(m1t )dt . (55)
N

t =1 u =1
m1 ≤ t m2 ≤ u 

For the eq. (53), we can rewrite the eq. (55) also as follow:

  a 
S a , q (α ; I ) = 4πiβ  ∑ ∫ ∑ λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a )e(− uβ )du  e( Nβ ) +
N

u =1
m1 ≤ N m2 ≤u  q 

N   a 
+ 8πβ 2 ∫  ∑ ∫ ∑ λ (m1 )λ (m2 )e (m1 − 2m2 )  − Cq (a )e(− uβ )du e(m1t )dt . (55b)
N

t =1 u =1
m1 ≤t m2 ≤u  q 

The factor of 8πβ 2 = −(4πiβ ) ⋅ (2πiβ ) and comes from the derivative of e(m1β ) . Arguing in a
similar manner as above in Theorem 1 and in Lemmas 5 and 6 we show the two terms in (55) are
small, which will complete the proof.

Lemma 5

   N2 
4πiβ  ∑ ∫ ∑ am 2 (m1 , N )e(− uβ )du  e( Nβ ) = O C − D  . (56)
N

u =1
m1 ≤ N m2 ≤u   log N

Proof. Arguing along the lines of Lemma 3, one shows the contribution from u ≤ N is bounded
by N log D N . For u ≥ N we apply the Siegel-Walfisz formula as in Lemma 3, giving a
contribution bounded by

15
   cq (a )u  u   cq (− 2a )N  N  
+ O C  − Cq (a )uN du <<
N
4β ∫  + O  ⋅ 
  φ (q )  log N   φ (q )
u= N C 
 log N  

N uN N3 β
<< β ∫ du << . (57)
u= N log C N log C N

log B N  N2 
As β ≤ , the above is O C − D  .
N  log N

Lemma 6

N    N2 
8πβ 2 ∫  ∑ ∫ ∑ am 2 (m1 , N )e(− uβ )du e(m1t )dt = O
N
 . (58)
t =1 u =1
 m1 ≤ t m2 ≤ u   log
C − 2 D
N 

Proof. Arguing as in Lemma 3, one shows that the contribution when t ≤ N or u ≤ N is

 N 
O  . We then apply the Siegel-Walfisz Theorem as before, and find the contribution
 log
C −2D
N 
when t , u ≥ N is

N N ut N 4β 2
<< 8β 2 ∫ ∫ dudt << . (59)
t= N u= N logC N logC N

log D N  N2 
As β ≤ , the above is O C − 2 D  . This complete the proof of Theorem 1.
N  log N

We note that, for the eq. (56) and (58), the eq. (55) can be rewritten also as follows:

 
S a , q (α ; I ) = 4πiβ  ∑ ∫ ∑ am 2 (m1 , N )e(− uβ )du  e( Nβ ) +
N

u =1
 m1 ≤ N m2 ≤ u 

N    N2   N2 
+ 8πβ 2 ∫  ∑ ∫ ∑ am2 (m1 , N )e(− uβ )du e(m1t )dt = O
N
C − D

 + O 
 C − 2 D
 . (59b)
t =1 u =1
 1
m ≤ t m 2 ≤ u   log N   log N 

With regard the integrals over the Major arcs, we first compute the integral of u ( x )e(− x ) over the
Major arcs and then use Theorem 1 to deduce the corresponding integral of FN ( x )e(− x ) .

By Theorem 1 we know for x ∈ Ma , q that

 a  N2 
FN ( x ) − Cq (a )u x −  << O C − 2 D  . (60)
 q  log N

16
  a
We now evaluate the integral of u  x − e(− x ) over Ma , q ; by Theorem 1 we then obtain the
 q
integral of FN ( x )e(− x ) over Ma , q . Remember that

u(x ) = ∑ e((m 1 − 2m2 )x ) . (61)

m1 , m 2 ≤ N

Theorem 2

 a  aN  N 
∫ u  α −  ⋅ e(− α )dα = e −  + O D
 . (62)
Ma ,q
 q  q 2  log N 

 1 1
We first determine the integral of u over all of − ,  , and then show that the integral of u ( x ) is
 2 2
Q
small if x > .
N

Lemma 7
1

∫ u (x )e(− x )dx = 2 + O(1) .

N
2
1 (63)
−
2

Proof.
1 1 1

∫ u (x )e(− x )dx = ∫ ∑ ∑ e((m

−
2
1
2
−
2
1
2 m1 ≤ N m 2 ≤ N
1 − 2m2 )x ) ⋅ e(− x )dx = ∑ ∑ ∫ e((m
m1 ≤ N m 2 ≤ N
−
2
1
2
1 − 2m2 − 1)x )dx . (64)

The integral is 1 if m1 − 2m2 − 1 = 0 and 0 otherwise. For m1 , m2 ∈ {1,..., N } , there are

N  N
 2  = 2 + O (1) solutions to m1 − 2m2 − 1 = 0 , which completes the proof.

Define

 1 Q Q Q 1 Q 
I1 = − + ,−  , I 2 =  , −  . (65)
 2 N N N 2 N 

The following bound is crucial in our investigations.

Lemma 8

for a ∈ {1,−2}.
1 1
For x ∈ I1 or I 2 , <<
1 − e(ax ) x

Lemma 9

 N 
∫ u ( x )e(− x )dx = O D
 . (66)
x∈I 1 ∪ I 2
 log N 

Proof. We have

17
 ∫ u (x )e(− x )dx = ∫ ∑ e((m
Ii Ii
m1 , m 2 ≤ N
1 − 2m2 − 1)x )dx = ∫
Ii
∑ e(m x ) ∑ e(− 2m x ) ⋅ e(− x )dx =
m1 ≤ N
1
m2 ≤ N
2

 e( x ) − e(( N + 1)x )   e(− 2 x ) − e(− 2( N + 1)x ) 

=∫ 
1 − e( x )  1 − e(− 2 x )  e(− x )dx (67)
Ii
  

because these are geometric series. By Lemma 8, we have

∫ u (x )e(− x )dx << ∫

22 N N
dx << = , (68)
Ii Ii xx Q log D N

which completes the proof of Lemma 9.

Lemma 10

( )
1 Q
+
∫ 2 N
1 Q u ( x )e(− x )dx = O log D N . (69)
x= −
2 N

Lemma 11
Q
 N 
u ( x )e(− x )dx =
N
∫ −
N
Q
N
2
+ O D
 . (70)
 log N 

Proof of Theorem 2. We have

 a a Q
+  a
Q
 a 
∫Ma ,q uα − q  ⋅ e(− α )dα = ∫aq − QN uα − q  ⋅ e(− α )dα = ∫−NQN u(β ) ⋅ e − q − β dβ =
  q N  

 a  QN  aN  N 
= e −  ∫ Q u (β )e(− β )dβ = e −  + O D
 . (71)
−
 q  N  q  2  log N 

 a
Note there are two factors in Theorem 2. The first, e −  , is an arithmetical factor which
 q
depends on which Major arc Ma, q we are in. The second factor is universal, and is the size of the
contribution.

An immediate consequence of Theorem 2 is

Theorem 3

 aN  N   
FN ( x )e(− x )dx = Cq (a )e −  + O
N
∫Ma ,q
 q 2
D
 + O C − 3 D  . (72)
 log N   log N

From Theorem 3 we immediately obtain the integral of FN ( x )e(− x ) over the Major arcs M :

Theorem 4

18
 log B N q
 aN  
FN ( x )e(− x )dx = Cq (a )e −  + O
N N
∫ ∑ ∑ D −2B
+ C −3D − 2 B
=
N 
M
q =1 a =1  q 2  log N log
( a , q ) =1

N  N N 
= ℑN + O +  , (73)
2  log
D−2B
N log C −3D −2 B
N 

log B N q
 a
where ℑN = ∑ ∑ Cq (a )e −  (74)
q =1 a =1  q
( a , q ) =1

is the truncated singular series for the Germain primes.

3. On some equations concerning the equivalence between the Goldbach’s Conjecture and the
Generalized Riemann Hypothesis [3]

We know the Goldbach’s conjecture: “Every even integer > 2 is the sum of two primes”. In 1922
Hardy and Littlewood guesstimated, via a heuristic based on the circle method, an asymptotic for
the number of representations of an even integer as the sum of two primes: Define

g (2 N ) = # {p, q prime : p + q = 2 N }.

Their conjecture is equivalent to g (2 N ) ≈ I (2 N ) where

 p − 1 2 N − 2
I (2 N ) = C2 ∏ 
dt
 ∫ (75)
p N  p − 2  2 log t log(2 N − t )
p >2

and C2 , the “twin prime constant”, is defined by

 1 
C2 = 2∏ 1 −  = 1.320323... (76)
p >2  ( p − 1)2 
Thence, the eq. (75) can be rewritten also as follows:
2 N −2
 1   p −1 
I (2 N ) = 2∏ 1 −
dt
2 ∏ 
    ∫ (77)
p>2  ( p − 1)  p N  p − 2  2 log t log(2 N − t )
p >2

and thence, we obtain:

2 N −2
 1   p −1 
g (2 N ) ≈ 2∏ 1 −
dt
2 ∏ 
    ∫ , (78)
p>2  ( p − 1)  p N  p − 2  2 log t log(2 N − t )
p >2

or

19
  p − 1 2 N − 2
g (2 N ) ≈ C2 ∏ 
dt
 ∫ . (78b)
p N  p − 2  2 log t log(2 N − t )
p >2

We believe that a better guesstimate for g (2 N ) is given by


I ∗ (2 N ) := I (2 N )1 −
4  (2 N / p )   .
∏ 1 − p − 2  
(79)
 2N p≥3 

Thence, for eq. (75) we can rewrite the eq. (79) also as follows:

I ∗ (2 N ) := C2 ∏ 
 p − 1 2 N − 2 dt  4  (2 N / p )   .
 ∫ 1 −
p N  p − 2  2 log t log(2 N − t ) 
 ∏ 1 − p − 2  
(79b)
2N p ≥3 
p >2

Indeed it could well be that

 N 
g (2 N ) = I ∗ (2 N ) + O log log N  . (80)
 log N 

Thence, for eq. (79b), we obtain the following equation:

g (2 N ) = C2 ∏ 
 p − 1 2 N − 2 dt  4  (2 N / p )   + O N 
 ∫ 1 −
p N  p − 2  2 log t log (2 N − t ) 
 ∏ 1 − p − 2  

 log N log log N  .
2N p ≥3   
p >2

(80b)

We introduce the function

G (2 N ) = ∑ log p log q . (81)

p+ q =2 N
p , q ( prime )

The analysis of Hardy and Littlewood suggests that G (2 N ) , plus some terms corresponding to
solutions of p k + q l = 2 N , should be very “well-approximated” by

 p −1 
J (2 N ) := C2 ∏   ⋅ 2 N , (82)
pN  p−2
p>2

and the approximation g (2 N ) ≈ I (2 N ) is then deduced by partial summation. (In fact we believe
( )
that G (2 N ) = J (2 N ) + O N 1 / 2 + o (1) .)

Theorem 1

The Riemann Hypothesis is equivalent to estimate

∑ (G(2 N ) − J (2 N )) << x
2N ≤x
3 / 2 + o (1)
. (83)

Theorem 2
20
The Riemann Hypothesis for Dirichlet L-functions L(s, χ ) , over all characters χ mod m which are
odd squarefree divisors of q , is equivalent to the estimate

2N ≤x
∑ (G(2 N ) − J (2 N )) << x 3 / 2 + o (1)
. (84)
2 N ≡ 2 ( mod q )

Theorem 3

The Riemann Hypothesis for Dirichlet L-functions L(s, χ ), χ mod q is equivalent to the conjectured
estimate

∑ G(2 N ) = φ (q ) ∑ G(2 N ) + O(x ).

1 1+ o (1)
(85)
2N ≤x 2N ≤x
q 2N

Let

E (2 N ) = ∑ log p log q . (86)

p k + ql =2 N
k , l ≥1
k + l ≥3

First note that

∑ log p log q ≤ log 2 N ⋅ ∑1 << N 1/ 3

log 2 N , (87)
pk +ql =2N pk ≤2N
k ≥3 k ≥3

and a similar argument works for l ≥ 3 . Also it is well-known that there are N o (1) pairs of integer
p, q with p 2 + q 2 = 2 N . Thus

E (2 N ) = 2 ∑ log p log q + O(N 1/ 3

log 2 N . (88) )
p+q2 =2N

Now, when we study solutions to p + q 2 = 2 N we find that l divides p if and only if

2 N ≡ q 2 mod l . Thus if (2 N / l ) = 0 or – 1 then l divides pq if and only if q ≡ 0 mod l . If
(2 N / l ) = 1 then there are 2 non-zero values of q mod l for which l divides p , and we also need
to count when l divides q . Therefore our factor is 2 if l = 2 , and

 2 + (2 N / l ) 
1 −  l / 2n
 l  times 1
 if 
(1 − 1 / l )2 (l − 1) / (l − 2 ) l 2 N , l ≥ 3

Now # {m, n > 0 : m + n 2 = 2 N } = 2 N + O(1) so we predict that

 (2 N / l )   p −1 
∑ log p log q ≈ ∏ 1 − C2 ∏   2 N , (89)
p +q 2 =2 N l ≥3  l − 2  p 2 N  p − 2 
p >2

and thus, after partial summation, that

21
  (2 N / l )   p −1  2N
2 ∑ 1 ≈ 4∏ 1 − C2 ∏   . (90)
p+q2 =2N l ≥3  l − 2  p 2 N  p − 2  log (2 N )
2

( p , q ) prime p >2

Subtracting this from I (2 N ) , we obtain the prediction I ∗ (2 N ) , as in (79). We can give the more
accurate prediction

 p − 1 2 N − 2   (2 N / p )  1  
I (2 N ) = C2 ∏ 
dt 1
∗
 ∫ 1 − ∏ 1 −  +  dt . (91)
p N  p − 2  2 log t log(2 N − t )
 p − 2  t 2N − t 
 p ≥3 
p >2

The explicit version of the Prime Number Theorem gives a formula of the form

xρ
∑ log p = x − ∑ ρ
( )
+ O log 2 x , (92)
p≤ x ρ
Im ρ ≤ x

where the sum is over zeros ρ of ζ (ρ ) = 0 with Re(ρ ) > 0 . In Littlewood’s famous paper he
investigates the sign of π ( x ) − Li ( x ) by a careful examination of a sum of the form
∑ρ : Im ρ ≤T
( )
Li x ρ , showing that this gets bigger than x1 / 2 −ε for certain values of x , and smaller
than – x1 / 2 −ε for other values of x . His method can easily be modified to show that the above
implies that

max ∑ log p − y = x B + o (1) (93)

y≤ x
p≤ y

where B = sup{Re ρ : ζ (ρ ) = 0} (note that 1 ≥ B ≥ 1 / 2 ). By partial summation it is not hard to

show that

x1+ ρ
∑ G (2 N ) = ∑ log p log q =
x2
2
−2 ∑
ρ (1 + ρ )
(
+ O x 2 B + o (1) ) (94)
2N ≤x p+ q≤ x ρ
Im ρ ≤ x

so that, by Littlewood’s method,

y2
max ∑ G (2 N ) − = x1+ B + o (1) . (95)
y≤ x 2
2N ≤ y

Therefore the Riemann Hypothesis (B = 1 / 2) is equivalent to the conjectured estimate

∑ G(2 N ) =
x2
2
( )
+ O x 3 / 2 + o (1) . (96)
2N ≤x

This implies Theorem 1 since

22
 µ 2 (d ) µ 2 (d )
∑ J (2n) = C2 ∑ 2n ∑ = 2C2 ∑ ∑ n
2n≤ x 2n≤ x dn
d (odd )
∏ ( p − 2)
pd d ≤x / 2
d (odd )
∏ ( p − 2)
pd n≤ x / 2
dn

µ 2 (d )
 x2  x2
= 2C2 ∑  + O ( x )  = + O( x log x ) . (97)
d ≤ x / 2 ∏ p d ( p − 2 )  8d
 2

d (odd )

Going further we note that for any coprime integers a, q ≥ 2

 
1  xρ 
∑ log p = 
φ (q ) 
x − ∑ χ (a ) ∑ 2
(
 + O log (qx ) ; (98) )
p≤ x χ ( mod q ) ρ : L ( ρ , χ )= 0 ρ 
p ≡ a ( mod q )  Im ρ ≤ x 

and thus

y B + o (1)
max
y≤ x
p≤ y
∑ log p −
φ (q )
=x q , (99)
p ≡ a ( mod q )

where Bq = sup{Re ρ : L(ρ , χ ) = 0} for some χ (mod q ) . R.C. Vaughan noted that by the same
methods but now using the above formula, we get a remarkable cancellation which leads to the
explicit formula

∑ G(2 N ) − φ (q ) ∑ G (2 N ) = φ (q ) χ (∑ ) χ (− 1)ρ (ρ∑λ ) cρ σ x ρ σ + O(x log (qx ))

1 1 + 2
, (100)
2N ≤x 2N ≤x mod q :L , = 0
q 2N χ ≠ χ0 σ : L (σ , χ )= 0
Im ρ , Im σ ≤ x

1
where cρ ,σ = ∫
1
(1 − t )ρ t σ −1dt is a constant depending only on ρ and σ . Thus Theorem 3 follows
0
ρ
since cρ ,σ ≤ (1 / ρ )∫ t σ −1dt = 1 / ρσ and has ∑ 1 / ρ << log 2 (qx ) . Thence, the eq. (100) can be
1

0 Im σ ≤ x

rewritten also as follows:

( )
1
G (2 N ) − G (2 N ) = ∑ χ (− 1)ρ :L (ρ∑ ∫ ρ (1 − t ) t dt ⋅x ρ +σ + O x log 2 (qx )
1 1 1
∑ ∑
ρ σ −1

2N ≤x φ (q ) 2 N ≤ x φ (q ) χ (mod q ) , λ )= 0 0
q 2N χ ≠ χ0 σ : L (σ , χ )= 0
Im ρ , Im σ ≤ x

(100b)

As in the proof of (97) we have

x2
∑ J (2n) =
2n≤ x 2φ (q )
+ O( x log x ) . (101)
q 2n

Now, Hardy and Littlewood showed that Generalized Riemann Hypothesis implies that

23
 ∑ G (2n) − J (2n ) << x 5 / 2 + o (1) . (102)
2

2n≤ x

We expect, as we saw in the precedent passages, that G (2n ) − J (2n ) << n1 / 2 + o (1) and so we
believe that

∑ G (2n ) − J (2n ) << x 2 +δ + o (1)

2
(103)
2n≤ x

for δ = 0 . This implies, by Cauchy’s inequality, that

 3+2δ + o (1) 
∑ G(2n) = ∑ J (2 n ) + O x
  (
 = x 2 / 2 + O x (3+δ ) / 2 + o (1) ) (104)
2n≤ x 2n≤ x  

by (97), which implies the Riemann Hypothesis if δ = 0 (as after (96) above); and implies that
ζ (ρ ) ≠ 0 if Re ρ > 3 / 4 if δ = 1 / 2 (that is, assuming Hardy and Littlewood’s (102)).

We find that (85) is too delicate to obtain the Riemann Hypothesis for L(s, χ ), χ (mod q ) from
(103). Instead we note that

 2 
x µ (m ) x ρ +1 
∑ G(2 N ) = cq  2 − 2m∑ ∑ χ (2)ρ :L(∑
p m ( p − 2 ) χ mod m

ρ , χ )= 0 ρ (ρ + 1) 
(105)
2N ≤x
2 N ≡ 2 ( mod q )
 q
m (odd )
∏ p >2 χ ( primitive )
 Im ρ ≤ x 

( 2 B q + o (1)
) , where c (2, q ) p( p − 2)
plus an error term O x q =
q
∏( pq
p odd ) ( p − 1)2
. As in (97) one can show that

x2
∑ J (2 N ) = cq + O(x log x ) , (106)
2N ≤x 2
2 N ≡ 2 ( mod q )

so that

∑ (G(2 N ) − J (2 N )) =O(x )
1+ C q + o (1)
(107)
2N ≤x
2 N ≡ 2 ( mod q )

where Cq = sup{Re ρ : L(ρ , χ ) = 0} for some χ mod m , where m q and m is odd and squarefree.
This implies Theorem 2. By the above we see that if (103) holds with δ = 0 then Cq = 1 / 2 and
thus the Riemann Hypothesis follows for L-functions with squarefree conductor.

4. On some equations concerning the p-adic strings and the zeta strings [4] [5] [6] [7].

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:
24
  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
, (108 – 111)

a = −α (s ) = −1 −
s
where h = 1 , T = 1 / π , and 2 , b = −α (t ) , c = −α (u ) with the condition
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
a −1 b −1
Qp
, (112)

... p
where 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
R
p∈P
Qp
, a ∈ Q× , b ∈ Q , (113)

what follows from

25
 χ v (ax 2 + bx )d v x = λv (a ) 2a v 2 χ v  −
−
1
 b2 
∫ 
 4a  , v = ∞,2,..., p... . (114)
Qv

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
x ' ', t ''
χv  −∫
x ', t '
 h t '
 , (115)

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

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

p ∈P
, x' ' , x' , λ ∈ Q , T ∈ Q× , (116)

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  − 
 24 4 8T 
. (117)

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:

λT  T ( x' '− x') 

2 3 2
+ [λ ( x' '+ x') − 2] +
1
K v ( x' ' , T ; x' ,0 ) = λv (− 8T ) 4T v 2 χ v  −
−
⇒
24 4 8T 
 
26
  ∞ cos πtxw' 
− πx 2 w '
 ∫0 cosh πx e dx  142
4 anti log πt 2 ⋅ 2
 e
−
4
w'
φ (itw ' )  t w'
⇒  
w '

  10 + 11 2   10 + 7 2  
log   + 
 


  4   4 
 . (117b)

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

ψ ( x ), x ∈ Z
( )
ψ A ( x ) = ψ ∞ ( x )∏ Ω x p =  ∞
p∈P 0, x ∈ Q \ Z , (118)

where
( )
Ω x p =1 x ≤1
if p and
Ω xp =0 ( )
x >1
if p . 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
R ζ (a ) , Q p 1 − p − a , (119)

B∞ (a, b ) = ∫ x ∞ 1 − x ∞ d ∞ x = Γ∞ (a )Γ∞ (b )Γ∞ (c )

a −1 b −1
R , (120)

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

a −1 b −1
Qp
, (121)

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:

Γ∞ (u )∏ Γp (u ) = 1 B∞ (a, b )∏ B p (a, b ) = 1
p∈P
, p∈P
, u ≠ 0,1, u = a, b, c, (122)

27
 B (a, b )
where a + b + c = 1 . We note that B∞ (a, b ) and p 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 
a
a −1 − a
R
 2  , (123)

ζ p (a ) =
1
1− p −1 ∫Q
p
( ) a −1
Ω x p x p dpx =
1
1 − p − a , Re a > 1 , (124)

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a )
p ∈P
, (125)

one obtains

ζ A (1 − a ) = ζ A (a ) , (126)

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 ) =
p ∈P ∫R exp − π(
x 2
x
a −1
∞
d)∞ x ⋅
1
1− p −1 ∫Q
p
( )
a −1
Ω x p x p dpx
. (126b)

(
Let us note that exp − πx and
2
)
Ω xp ( )
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
p ∈P
, (127)

whose the Fourier transform

28
 ∏ Ω (k )
1
ψ A (k ) = ∫ χ A (kx )ψ A (x ) = 2 e −πk
2
4 ∞
p p
p ∈P
(128)

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

( )
a
 a  −2
Φ A (a ) = ∫ψ A ( x ) x d x = ∫ ψ ∞ ( x ) x  π ζ (a )
1
d ∞ x∏
× a −1 a −1
−1 ∫Q
Ω = Γ
a
x x d x 2
p∈P 1 − p
A p
R p
p
2 (129)

and the same for ψ A (k ) . Then according to the Tate formula one obtains (126).

The exact tree-level Lagrangian for effective scalar field ϕ which describes open p-adic string
tachyon is

1 p2  1 − 2 


1
Lp = 2 − ϕp ϕ + ϕ p +1 
g p −1  2 p +1  , (130)

= −∂ t2 + ∇ 2
where p is any prime number, 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 
n ≥1 n + 1
2 n
n ≥1 n ≥1 n g  2 n ≥1  . (131)

Recall that the Riemann zeta function is defined as

ζ (s ) = ∑
1 1
=∏ −s
p 1− p
s
n ≥1 n
, s = σ + iτ , σ > 1 . (132)

Employing usual expansion for the logarithmic function and definition (132) we can rewrite (131)
in the form

29
 1 1 
 
L=− 2 
φζ  φ + φ + ln (1 − φ )
g 2  2   , (133)

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


 ixk  k 2 ~
ζ  φ (x ) = ∫  2 φ (k )dk
1
e ζ  − r
2 (2π )D ,
− k 2 = k02 − k 2 > 2 + ε , (134)

φ (k ) = ∫ e(−ikx )φ ( x )dx
~
where 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
eixk ζ  − φ (k )dk =
1
ζ  φ =
(2π )D ∫
r
2 k 02 − k 2 > 2 + ε
 2 1−φ
(135)

which has an evident solution φ = 0 .

For the case of time dependent spatially homogeneous solutions, we have the following equation
of motion for the zeta string

 − ∂ t2   k2 ~ φ (t )
φ (t ) = e − ik 0 tζ  0 φ (k0 )dk0 =
1
ζ 
 2  (2π ) ∫k 0 > 2 +ε
 2 1 − φ (t )
. (136)

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

30
 n ( n −1)

 ixk  k2 ~
e ζ  − φ (k )dk = ∑θ 2 φ n
1
ζ  φ = D ∫
2 (2π )  2 n ≥1
, (137)

 n 2 n(n − 1) n (n2−1) −1 n +1 


ζ  θ =
1 ixk 
ζ
∫  4 
e  −
k2 ~
θ (k )dk = ∑ θ +
2(n + 1)
θ ( )
φ −1 
4 (2π )D n ≥1   , (138)

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

The exact tree-level Lagrangian of effective scalar field ϕ , which describes open p-adic string
tachyon, is:

m Dp p 2  1 − 2 m 2p 


1
Lp = 2 − ϕp ϕ+ ϕ p +1 
g p p −1  2 p +1 
  , (139)

= −∂ t2 + ∇ 2
where p is any prime number, is the D-dimensional d’Alambertian and we adopt
metric with signature (− +... + ) , as above. Now, we want to introduce a model which incorporates
all the above string Lagrangians (139) with p replaced by n ∈ N . Thence, we take the sum of all
L
Lagrangians n in the form

mnD n 2  1 − 2 mn2 1 n +1 


+∞ +∞
L = ∑ Cn Ln = ∑ Cn 2 − φn φ+ φ 
n =1 n =1 gn n − 1 2 n +1 
 , (140)

whose explicit realization depends on particular choice of coefficients Cn , masses mn and coupling
constants g n .

Now, we consider the following case

n −1
Cn =
n 2 + h , (141)

31
where h is a real number. The corresponding Lagrangian reads

mD  1 +∞ −
2 − h +∞
n − h n +1 
Lh = 2 − φ ∑ n 2 m φ + ∑ φ 
g  2 n =1 n =1 n + 1  (142)

and it depends on parameter h . According to the Euler product formula one can write

+∞

− −h 1
∑n 2m2
=∏

n =1 − −h
1− p
p 2m2
. (143)

Recall that standard definition of the Riemann zeta function is

+∞
ζ (s ) = ∑
1 1
=∏ −s
p 1− p
s
n =1 n
, s = σ + iτ , σ > 1 , (144)

which has analytic continuation to the entire complex s plane, excluding the point s = 1 , where it
has a simple pole with residue 1. Employing definition (144) we can rewrite (142) in the form

mD  1 
 +∞
n − h n +1 
Lh =
g2
−
 2 φζ 
 2m
2
+ h 

φ + ∑
n =1 n + 1
φ 
  . (145)



ζ 2
+ h
Here  2m  acts as a pseudodifferential operator

32
 
 ixk  k2 ~
ζ + h φ ( x ) =
1
D ∫
e ζ 
 − + h φ (k )dk
 2m
2
 (2π )  2m
2
 , (146)

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

We consider Lagrangian (145) with analytic continuations of the zeta function and the power
n − h n +1
∑ n + 1φ
series , i.e.

mD  1 
 +∞
n − h n +1 
Lh =
g2
− φζ
 2  2m 2

+ h 

φ + AC ∑ +
φ 
 n =1 n 1  , (147)

where AC denotes analytic continuation.

Potential of the above zeta scalar field (147) is equal to − Lh at
= 0 , i.e.

mD  φ 2 +∞
n − h n +1 
Vh (φ ) =  ζ (h ) − AC ∑ φ 
g 2  2 n =1 n + 1  , (148)

where h ≠ 1 since ζ (1) = ∞ . The term with ζ -function vanishes at h = −2,−4,−6,... . The equation
of motion in differential and integral form is

+∞


ζ  2 + h φ = AC ∑ n − hϕ n
 2m  n =1 , (149)

+∞
ixk  k2 ~
1
e ζ
∫R D  2m2 
 − + h φ (k )dk = AC ∑ n − hφ n
(2π )D n =1
, (150)

respectively.

33
Now, we consider five values of h , which seem to be the most interesting, regarding the
Lagrangian (147): h = 0, h = ±1, and h = ±2 . For h = −2 , the corresponding equation of motion
now read:


  k2 ~ φ (φ + 1)
∫R D e ζ  − 2m 2 − 2 φ (k )dk = (1 − φ )3
1
ζ  2 − 2 φ = ixk

 2m  (2π )D . (151)

This equation has two trivial solutions: φ ( x ) = 0 and φ ( x ) = −1 . Solution φ ( x ) = −1 can be also

shown taking (k ) = −δ (k )(2π ) and ζ (− 2) = 0 in (151).

~
φ D

For h = −1 , the corresponding equation of motion is:


 ixk  k2 ~ φ
∫R D  2m2 − 1φ (k )dk = (1 − φ )2
1
ζ − 1φ = e ζ  −
 2m
2
 (2π )D . (152)

ζ (− 1) = −
1
where 12 .

The equation of motion (152) has a constant trivial solution only for φ ( x ) = 0 .

For h = 0 , the equation of motion is


 ixk  k2 ~ φ
∫R D  2m2 φ (k )dk = 1 − φ
1
ζ 2 
φ= e ζ  −
 2m  (2π )D . (153)

It has two solutions: φ = 0 and φ = 3 . The solution φ = 3 follows from the Taylor expansion of the
Riemann zeta function operator

ζ (n ) (0 ) 

n


ζ  2  = ζ (0 ) + ∑  
 2m  n ≥1 n!  2m 2  , (154)

as well as from φ (k ) = (2π ) 3δ (k ) .

~ D

34
For h = 1 , the equation of motion is:

ixk  k2 ~
∫R D  2m2 + 1φ (k )dk = − 2 ln(1 − φ )
1 1
ζ  −
2
e
(2π )D , (155)

where ζ (1) = ∞ gives V1 (φ ) = ∞ .

In conclusion, for h = 2 , we have the following equation of motion:

φ ln (1 − w)
ixk  ~ 2
k2
1
e ζ
∫R D  2m 2 
 − + 2 φ (k )dk = − ∫0 2w dw
(2π )D . (156)

Since holds equality

ln (1 − w)
dw = ∑ n =1 2 = ζ (2 )
1 ∞ 1
−∫
0 w n

one has trivial solution φ = 1 in (156).

n2 − 1
Cn =
Now, we want to analyze the following case: n 2 . In this case, from the Lagrangian (140),
we obtain:

mD  1  
 
 φ2 
L=  − φ ζ
  2 − 1 + ζ  2  φ + 
g2  2   2m   2m  1 − φ  . (157)

The corresponding potential is:

m D 31 − 7φ 2
V (φ ) = − φ
g 24(1 − φ ) . (158)

35
We note that 7 and 31 are prime natural numbers, i.e. 6n ± 1 with n =1 and 5, with 1 and 5 that
are Fibonacci’s numbers. Furthermore, the number 24 is related to the Ramanujan function that
has 24 “modes” that correspond to the physical vibrations of a bosonic string. Thence, we obtain:

 ∞ cos πtxw' 
−πx 2 w '
 ∫0 cosh πx e dx  142
4 anti log πt 2 ⋅ 2
 e
−
4
w'
φ (itw ' )  t w'
⇒  
w'

  10 + 11 2   10 + 7 2  
m D 31 − 7φ 2 log   +  
V (φ ) = − φ   
g 24(1 − φ )   4   4 
 . (158b)

The equation of motion is:

 
 
 [
φ (φ − 1) + 1
2
]
ζ  2m 2 − 1 + ζ  2m 2 φ = (φ − 1)2
     . (159)

Its weak field approximation is:

 
 
 
ζ  2m 2 − 1 + ζ  2m 2  − 2φ = 0
      , (160)

which implies condition on the mass spectrum

 M2   M2 
ζ  2 − 1 + ζ  2  = 2
 2m   2m  . (161)

From (161) it follows one solution for M > 0 at M ≈ 2.79m and many tachyon solutions when
2 2 2

M 2 < −38m 2 .

36
 5 −1 5 +1
φ= Φ=
We note that the number 2.79 is connected with 2 and 2 , i.e. the “aurea”
section and the “aurea” ratio. Indeed, we have that:

2
 5 + 1 1  5 −1
  + 2 
 2  2  2  = 2,772542 ≅ 2,78
    .

Furthermore, we have also that:

(Φ )14 / 7 + (Φ )−25 / 7 = 2,618033989 + 0,179314566 = 2,79734

With regard the extension by ordinary Lagrangian, we have the Lagrangian, potential, equation of
n2 − 1
Cn =
motion and mass spectrum condition that, when n 2 , are:

mD φ 

 
  φ
2
φ2 
L=   2 − ζ  − 1  − ζ   − 1φ + ln φ 2
+ 
g2 2 m  2m
2

2
 2m   2 1−φ 
, (162)

mD φ 2  1 
V (φ ) = ζ (− 1) + ζ (0 ) + 1 − ln φ 2 −
2
g 2  1 − φ  , (163)

 
 

 2φ − φ 2
ζ 
 2m 2 − 1  + ζ   − + 1φ = φ ln φ 2
+ φ +
  
2
 2m  m
2
 (1 − φ )2 , (164)

 M2   M2  M2
ζ  2
− 1
 + ζ  2  = 2
 2m   2m  m . (165)

In addition to many tachyon solutions, equation (165) has two solutions with positive mass:
M 2 ≈ 2.67 m 2 and M 2 ≈ 4.66m 2 .

We note also here, that the numbers 2.67 and 4.66 are related to the “aureo” numbers. Indeed,
we have that:

37
 2
 5 + 1 1  5 −1
  +  
 2  2 ⋅ 5  2  ≅ 2.6798
    ,
2
 5 + 1  5 + 1 1  5 + 1
     
 2  +  2  + 2 2  2  ≅ 4.64057
      .

Furthermore, we have also that:

(Φ )14 / 7 + (Φ )−41 / 7 = 2,618033989 + 0,059693843 = 2,6777278 ;

(Φ )22 / 7 + (Φ )−30 / 7 = 4,537517342 + 0,1271565635 = 4,6646738 .

n −1
Cn = µ (n )
Now, we describe the case of n 2 . Here µ (n ) is the Mobius function, which is defined
for all positive integers and has values 1, 0, – 1 depending on factorization of n into prime
numbers p . It is defined as follows:

0, n = p 2 m
 
µ (n ) = (− 1)k , n = p1 p2 ... pk , pi ≠ p j
1, 
 n = 1, (k = 0 ). (166)

The corresponding Lagrangian is

 
m D  1 + ∞ µ (n ) +∞
µ (n ) n +1 
Lµ = C0 L0 + 2 − φ ∑
φ + ∑ φ
g  2 n =1 2 m 2 n =1 n + 1 
 n  (167)

Recall that the inverse Riemann zeta function can be defined by

38
 1 +∞
µ (n )
=∑ s
ζ (s ) n =1 n , s = σ + it , σ > 1 . (168)

Now (167) can be rewritten as

 
 D 
φ + ∫ M (φ )dφ 
m 1 1 ∞
Lµ = C0 L0 + 2 − φ
g  2 
 0

ζ 2
  2 m  
, (169)

M (φ ) = ∑ n =1 µ (n )φ n = φ − φ 2 − φ 3 − φ 5 + φ 6 − φ 7 + φ 10 − φ 11 − ...
+∞

where The corresponding potential,

equation of motion and mass spectrum formula, respectively, are:

Vµ (φ ) = − Lµ (
= 0) =
m D  C0 2
2 
g 2
( φ
) 
φ 1 − ln φ 2 − φ 2 − ∫ M (φ )dφ 
0
 , (170)

φ − M (φ ) − C0 2 φ − 2C0φ ln φ = 0
1

 m
ζ 2
 2m  , (171)

1 M2
− C 0 + 2C0 − 1 = 0
 M2  m2
ζ  2 
 2m  φ << 1
, , (172)

r2
where usual relativistic kinematic relation k 2
= − k 2
0 + k = − M 2 is used.

Now, we take the pure numbers concerning the eqs. (161) and (165). They are: 2.79, 2.67 and
5 +1
Φ=
4.66. We note that all the numbers are related with 2 , thence with the aurea ratio, by
the following expressions:

2,79 ≅ (Φ ) 2,67 ≅ (Φ ) + (Φ ) 4,66 ≅ (Φ ) + (Φ )

15 / 7 13 / 7 −21 / 7 22 / 7 −30 / 7
; ; . (173)
39
 5. Mathematical connections

We take the eqs. (7) and (10b). We have the following expression:

1−δ 2 2
1  sin(π ( N + 1)α )   sin(π ( N + 1)α ) 
1
1 f N ( z )k dz
rk (n) = ∫ ⇒ r ( n ) = ∫   d α = 2 ∫   dα . (174)
2π i n +1
2π i 0  sin(πα ) πα
2
z  δ  sin( ) 

We have the following possible mathematical connection with the eq. (126b) concerning the adelic strings:

1−δ 2 2
1  sin(π ( N + 1)α )   sin(π ( N + 1)α ) 
1
1 f N ( z )k dz
rk (n) = ∫ ⇒ r2 (n) = ∫   dα = 2 ∫   dα ⇒
2π i z n +1
2π i 0  sin(πα )  δ  sin(πα ) 

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) =
p ∈P ∫R exp − (
πx 2
x
a −1
∞
)d ∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p dpx
⇒ p
. (175)

We note that also the eqs. (22), (39) and (50) can be connected with the (126b), as follows:
ξ '( q , a )
q
µ (q) 2 a
RM (n) ≈ ∑∑ *
∫ T (η ) 2 e(− n( + η ))dη =
q ≤ P a =1 −ξ ( q , a )
ϕ (q) 2
q

ξ '( q , a )
µ (q)2 q * a
=∑ 2 ∑
e(− n ) ∫ T (η ) 2 e(− nη )dη ⇒
q ≤ P ϕ ( q ) a =1 q −ξ ( q , a )

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) =
p ∈P ∫R exp − (
πx 2
x
a −1
∞
)d ∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p dpx
⇒ p
, (176)

  
S a , q (α ) = ∑   ∑ a (m , N )e(− 2 Nβ )e(m β ) +
m2 1 1
m1 ≤ N  m2 ≤ N  

   
∑ 4πiβ ∫ ∑ am2 (m1 , N ) e(− uβ )du  e(m1β ) ⇒
N
+ 
u =1
m1 ≤ N  m2 ≤ u  

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) =
p ∈P ∫R exp − (
πx 2
x
a −1
∞
)d ∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p dpx
⇒ p
, (177)

40
    N2 
( ) ( )
N
2πiβ ∫
t =0
∑  ∑ m2 1  a m , N e t β dt = O  C − D  ⇒
m1 ≤ t  m 2 ≤ N   log N

 N 3β   N2 
( )
N
⇒β∫
t= N
∑ ∑ m2 1 a m , N dt = O  
 log C N  = O  C − D  ⇒
m1 ≤ t m 2 ≤ N    log N

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) =
p ∈P ∫R exp (
− πx 2
x )
a −1
∞
d ∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p dpx
⇒ p
. (178)

While the eq. (59b) can be related further that with the eq. (126b) also with the Ramanujan
modular identity concerning the physical vibrations of the superstrings, i.e. the number 8, that is
also a Fibonacci’s number. Thence, we have that

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

 
⇒ S a , q (α ; I ) = 4πiβ  ∑ ∫ ∑ am2 (m1 , N )e(− uβ )du  e( Nβ ) +
N

u =1
 m1 ≤ N m2 ≤ u 

N    N2   N2 
+ 8πβ 2 ∫  ∑ ∫ ∑ am2 (m1 , N )e(− uβ )du e(m1t )dt = O C − D  + O C − 2 D  ⇒
N

t =1 u =1
m1 ≤ t m2 ≤ u   log N   log N

 ∞ 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 '
. (179)
3   10 + 11 2   10 + 7 2  
log   + 
 


  4   4 


Also the eq. (73) can be related with the eq. (126b), thence we obtain the following expression:

log B N q
 aN  
FN ( x )e(− x )dx = Cq (a )e −  + O
N N
∫ ∑ ∑ D −2B
+ C −3D − 2 B
=
N 
M
q =1 a =1  q 2  log N log
( a , q ) =1

N  N N 
= ℑN + O + ⇒
2  log
D−2B
N log C −3D − 2 B
N 

ζ A (a ) = ζ ∞ (a )∏ ζ p (a ) = ζ ∞ (a )ζ (a ) =
p ∈P ∫R exp (
− πx 2
x )
a −1
∞
d ∞ x ⋅
1
1− p −1 ∫Q
( ) a −1
Ω x p x p dpx
⇒ p
. (180)

41
Now, with regards the mathematical connections with the zeta strings, we have that the eq. (78b)
can be related with the eq. (136), that is the equation of motion for the zeta string concerning the
case of time dependent spatially homogeneous solutions. Thence, we obtain:

 p − 1 2 N − 2
g (2 N ) ≈ C2 ∏ 
dt
 ∫ ⇒
p N  p − 2  2 log t log(2 N − t )
p >2

 − ∂ t2   k2 ~ φ (t )
φ (t ) = e − ik 0 tζ  0 φ (k0 )dk0 =
1
ζ 
⇒  2  (2π ) ∫k 0 > 2 +ε
 2 1 − φ (t )
. (181)

With regard the eq. (80b), it can be related with the eq. (138) i.e. the equation of motion
concerning the closed scalar zeta strings:

g (2 N ) = C2 ∏ 
 p − 1 2 N − 2 dt  4  (2 N / p )   + O N 
 ∫ 1 −
p N  p − 2  2 log t log (2 N − t ) 
 ∏ 1 − p − 2    log N log log N ⇒

2N p≥3   
p >2

 n 2 n(n − 1) n (n2−1) −1 n +1 


ζ  θ =
1 ixk 
ζ
∫  4 
e  −
k2 ~
θ (k )dk = ∑ θ +
2(n + 1)
θ (
φ −1  )
⇒
4 (2π )D n ≥1   . (182)

With regard the eq. (91), it can be related with the eq. (138) and with the eq. (156), thence we
obtain the following expressions:

 p − 1 2 N − 2   (2 N / p )  1  
I ∗ (2 N ) = C2 ∏ 
dt 1
 ∫ 1 − ∏ 1 −  +  dt ⇒
p N  p − 2  2 log t log(2 N − t )
 p − 2  t 2N − t 
 p ≥3 
p >2

 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  ( )
⇒
4 (2π )  4 n ≥1   , (183)

 p − 1 2 N − 2   (2 N / p )  1  
I ∗ (2 N ) = C2 ∏ 
dt 1
 ∫ 1 − ∏ 1 −  +  dt ⇒
p N  p − 2  2 log t log(2 N − t )
 p − 2  t 2N − t 
 p ≥3 
p >2

φ ln (1 − w)
ixk  ~ 2
k2
⇒
1
e ζ
∫R D  2m 2 
 − + 2 φ (k )dk = − ∫0 2w dw . (184)
(2π )D
In conclusion, the eq. (100b) can be related with the eq. (156) and we obtain the following
mathematical connection:

42
 ( )
1
G (2 N ) − G (2 N ) = ∑ χ (− 1)ρ :L (ρ∑ ∫ ρ (1 − t ) t dt ⋅x ρ +σ + O x log 2 (qx ) ⇒
1 1 1
∑ ∑
ρ σ −1

2N ≤x φ (q ) 2 N ≤ x φ (q ) χ (mod q ) , λ )= 0 0
q 2N χ ≠ χ0 σ : L (σ , χ )= 0
Im ρ , Im σ ≤ x

φ ln (1 − w)
ixk  ~ 2
k2
⇒
1
e ζ
∫R D  2m 2 
 − + 2 φ (k )dk = − ∫0 2w dw . (185)
(2π )D
We want to evidence, also in this paper, the fundamental connection between π and
5 −1
φ= , i.e. the Aurea ratio by the simple formula
2

arccos φ = 0,2879π . (186)

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.

References

[1] Rosario Turco: “ Il Metodo del Cerchio” – Gruppo Eratostene –

http://www.gruppoeratostene.com/articoli/Cerchio.pdf.

[2] Steven J. Miller and Ramin Takloo-Bighash: “The Circle Method” – pg 1- 73 ;July 14, 2004 –
Chapter of the book: “An Invitation to Modern Number Theory” Princeton University Press ISBN:
9780691120607 - March 2006.

[3] Andrew Granville: “Refinements of Goldbach’s Conjecture, and the Generalized Riemann
Hypothesis” – Functiones et Approximatio XXXVII (2007), 7 – 21.

[4] Branko Dragovich: “Adeles in Mathematical Physics” – arXiv:0707.3876v1 [hep-th]– 26 Jul

2007.

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

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

– 25 Apr 2008.

[7] Branko Dragovich: “Some Lagrangians with Zeta Function Nonlocality” – arXiv:0805.0403 v1 –
[hep-th] – 4 May 2008.

43