On some mathematical connections between Fermat’s Last Theorem, Modular Functions,

Modular Elliptic Curves and some sector of String Theory

Michele Nardelli 1, 2

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

This paper is fundamentally a review, a thesis, of principal results obtained in some sectors of
Number Theory and String Theory of various authoritative theoretical physicists and
mathematicians.
Precisely, we have described some mathematical results regarding the Fermat’s Last Theorem, the
Mellin transform, the Riemann zeta function, the Ramanujan’s modular equations, how primes and
adeles are related to the Riemann zeta functions and the p-adic and adelic string theory.
Furthermore, we show that also the fundamental relationship concerning the Palumbo-Nardelli
model (a general relationship that links bosonic string action and superstring action, i.e. bosonic and
fermionic strings in all natural systems), can be related with some equations regarding the p-adic
(adelic) string sector.
Thence, in conclusion, we have described some new interesting connections that are been obtained
between String Theory and Number Theory, with regard the arguments above mentioned.
In the Chapters 1 and 2, we have described the mathematics concerning the Fermat’s Last
Theorem, precisely the Wiles approach in the Chapter 1 and further mathematical aspects
concerning the Fermat’s Last Theorem, precisely the modular forms, the Euler products, the
Shimura map and the automorphic L-functions in the Chapter 2. Furthermore. In this chapter, we
have described also some mathematical applications of the Mellin transform, only mentioned in the
Chapter 1, the zeta-function quantum field theory and the quantum L-functions.
In the Chapter 3, we have described how primes and adeles are related to the Riemann zeta
function, precisely the Connes approach. In the Chapter 4, we have described the p-adic and adelic
strings, precisely the open and closed p-adic strings, the adelic strings, the solitonic q-branes of p-
adic string theory and the open and closed scalar zeta strings.
In the Chapter 5, we have described some correlations obtained between some solutions in string
theory, Riemann zeta function and Palumbo-Nardelli model. Precisely, we have showed the
cosmological solutions from the D3/D7 system, the classification and stability of cosmological
solutions, the solution applied to ten dimensional IIB supergravity, the connections with some
equations concerning the Riemann zeta function, the Palumbo-Nardelli model and the Ramanujan’s
identities. Furthermore, we have described the interactions between intersecting D-branes and the
general action and equations of motion for a probe D3-brane moving through a type IIB
supergravity background. Finally, in the Chapter 6, we have showed the connections between the
equations of the various chapters.

1
 Introduzione e riassunto

L’ultimo teorema di Fermat è una generalizzazione dell’equazione diofantea a 2 + b 2 = c 2 . Già gli

antichi Greci ed i Babilonesi sapevano che questa equazione ha delle soluzioni intere, come (3, 4, 5)
(3 2 + 4 2 = 5 2 ) o (5, 12, 13) (5 2 + 12 2 = 13 2 ) . Queste soluzioni sono conosciute come “terne
pitagoriche” e ne esistono infinite, anche escludendo le soluzioni banali per cui a, b e c hanno un
divisore in comune e quelle ancor più banali in cui almeno uno dei numeri è uguale a zero.
Secondo l’ultimo teorema di Fermat, non esistono soluzioni intere positive quando l’esponente 2 è
sostituito da un numero intero maggiore. Il teorema è particolarmente noto per la sua correlazione
con molti argomenti matematici che apparentemente non hanno nulla a che vedere con la Teoria dei
Numeri. Inoltre, la ricerca di una dimostrazione è stata all’origine dello sviluppo di importanti aree
della matematica, anche legate a moderni settori della fisica teorica, quali ad esempio la Teoria
delle Stringhe.
L’ultimo teorema di Fermat può essere dimostrato per n = 4 e nel caso in cui n è un numero primo:
se infatti si trova una soluzione a kp + b kp = c kp , si ottiene immediatamente una soluzione
(a ) + (b ) = (c )
k p k p k p
. Nel corso degli anni il teorema venne dimostrato per un numero sempre
maggiore di esponenti speciali n, ma il caso generale rimaneva evasivo. Il caso n = 5 è stato
dimostrato da Dirichlet e Legendre nel 1825 ed il caso n = 7 da Gabriel Lamé nel 1839. Nel 1983
G. Faltings dimostrò la congettura di Mordell, che implica che per ogni n > 2, c’è al massimo un
numero finito di interi “co-primi” a, b e c con a n + b n = c n . (In matematica, gli interi a e b si
dicono “co-primi” o “primi tra loro” se e solo se essi non hanno nessun divisore comune eccetto 1 e
-1, o, equivalentemente, se il loro massimo comune divisore è 1).
Utilizzando i sofisticati strumenti della geometria algebrica (in particolare curve ellittiche e forme
modulari), della teoria di Galois e dell’algebra di Hecke, il matematico di Cambridge Andrew John
Wiles, dell’Università di Princeton, con l’aiuto del suo primo studente, Richard Taylor, diede una
dimostrazione dell’ultimo teorema di Fermat, pubblicata nel 1995 nella rivista specialistica “Annals
of Mathematics”.
Nel 1986, Ken Ribet aveva dimostrato la “Congettura Epsilon” di Gerhard Frey secondo la quale
ogni contro-esempio a n + b n = c n all’ultimo teorema di Fermat avrebbe prodotto una curva ellittica
definita come: ( )( )
y 2 = x ⋅ x − a n ⋅ x + b n , che fornirebbe un contro-esempio alla “Congettura di
Taniyama-Shimura”. Quest’ultima congettura propone un collegamento profondo fra le curve
ellittiche e le forme modulari. Wiles e Taylor hanno stabilito un caso speciale della Congettura di
Taniyama-Shimura sufficiente per escludere tali contro-esempi in seguito all’ultimo teorema di
Fermat. In pratica, la dimostrazione che le curve ellittiche semistabili sui razionali sono modulari,
rappresenta una forma ridotta della Congettura di Taniyama-Shimura che tuttavia è sufficiente per
provare l’ultimo teorema di Fermat.
Le curve ellittiche sono molto importanti nella Teoria dei Numeri e ne costituiscono il maggior
campo di ricerca attuale. Nel campo delle curve ellittiche, i “numeri p-adici” sono conosciuti come
“numeri l-adici”, a causa dei lavori di Jean-Pierre Serre. Il numero primo p è spesso riservato per
l’aritmetica modulare di queste curve.
Il sistema dei numeri p-adici è stato descritto per la prima volta da Kurt Hensel nel 1897. Per ogni
numero primo p, il sistema dei numeri p-adici estende l’aritmetica dei numeri razionali in modo
differente rispetto l’estensione verso i numeri reali e complessi. L’uso principale di questo
strumento viene fatto nella Teoria dei Numeri. L’estensione è ottenuta da un’interpretazione
alternativa del concetto di valore assoluto. Il motivo della creazione dei numeri p-adici era il
tentativo di introdurre il concetto e le tecniche delle “serie di potenze” nel campo della Teoria dei
Numeri. Più concretamente per un dato numero primo p, il campo Q p dei numeri p-adici è
un’estensione dei numeri razionali. Se tutti i campi Q p vengono considerati collettivamente, si

2
arriva al “principio locale-globale” di Helmut Hasse, il quale, a grandi linee, afferma che certe
equazioni possono essere risolte nell’insieme dei numeri razionali se e solo se possono essere risolte
negli insiemi dei numeri reali e dei numeri p-adici per ogni p. Il campo Q p possiede una topologia
derivata da una metrica, che è, a sua volta, derivata da una stima alternativa dei numeri razionali.
Questa metrica è “completa”, nel senso che ogni serie di Cauchy converge.
Scopo del presente lavoro è quello di evidenziare le connessioni ottenute tra la matematica inerente
la dimostrazione dell’ultimo teorema di Fermat ed alcuni settori della Teoria di Stringa,
precisamente la supersimmetria p-adica e adelica in teoria di stringa.
I settori inerenti la dimostrazione dell’ultimo teorema di Fermat, riguardano quelle funzioni
chiamate L p-adiche connesse alla funzione zeta di Riemann, quale estensione analitica al piano
complesso della serie di Dirichlet. Tali funzioni sono strettamente correlate sia ai numeri primi, sia
alla funzione zeta, i cui teoremi sono già stati connessi matematicamente con la teoria di stringa in
alcuni precedenti lavori.
Quindi, per concludere, anche dalla matematica che riguarda l’ultimo teorema di Fermat è possibile
ottenere, come vedremo nel corso del lavoro, ulteriori connessioni tra Teoria di Stringa (p-adic
string theory), Numeri Primi, Funzione zeta di Riemann (numeri p-adici, funzioni L p-adiche) e
Serie di Fibonacci (quindi identità e funzioni di Ramanujan), che, a loro volta, verranno correlate
anche al modello Palumbo-Nardelli.

Chapter 1.

The mathematics concerning the Fermat’s Last Theorem

1.1 The Wiles approach.[1]

An elliptic curve over Q is said to be modular if it has a finite covering by a modular curve of the
form X 0 ( N ) . Any such elliptic curve has the property that its Hasse-Weil zeta function has an
analytic continuation and satisfies a functional equation of the standard type. If an elliptic curve
over Q with a given j-invariant is modular then it is easy to see that all elliptic curves with the same
j-invariant are modular. A well-known conjecture which grew out of the work of Shimura and
Taniyama in the 1950’s and 1960’s asserts that every elliptic curve over Q is modular.
In 1985 Frey made the remarkable observation that this conjecture should imply Fermat’s Last
Theorem. The Wiles approach to the study of elliptic curves is via their associated Galois
representations. Suppose that ρ p is the representation of Gal (Q / Q ) on the p-division points of an
elliptic curve over Q, and suppose that ρ 3 is irreducible. The choice of 3 is critical because a
crucial theorem of Langlands and Tunnell shows that if ρ 3 is irreducible then it is also modular.
Thence, under the hypothesis that ρ 3 is semistable at 3, together with some milder restrictions on
the ramification of ρ 3 at the other primes, every suitable lifting of ρ 3 is modular. Furthermore,
Wiles has obtained that E is modular if and only if the associated 3-adic representation is modular.
The key development in the proof is a new and surprising link between two strong but distinct
traditions in number theory, the relationship between Galois representations and modular forms on
the one hand and the interpretation of special values of L-functions on the other.
The restriction that ρ 3 be irreducible at 3 is bypassed by means of an intriguing argument with
families of elliptic curves which share a common ρ 5 . Using this, we complete the proof that all
semistable elliptic curves are modular. In particular, this yields to the proof of Fermat’s Last
Theorem.
Now we present the methods and results in more detail.

3
Let f be an eigenform associated to the congruence subgroup Γ1 ( N ) of SL2 (Z ) of weight k ≥ 2 and
character χ . Thus if Tn is the Hecke operator associated to an integer n there is an algebraic integer
c(n, f ) such that Tn f = c(n, f ) f for each n. We let K f be the number field generated over Q by
the {c(n, f )} together with the values of χ and let Ο f be its ring of integers. For any prime λ of
Ο f let Ο f ,λ be the completion of Ο f at λ . The following theorem is due to Eichler and Shimura
(for k > 2).

THEOREM 1.

For each prime p ∈ Z and each prime λ p of Ο f there is a continuous representation

ρ f ,λ : Gal (Q / Q ) → GL2 (Ο f ,λ ) (1)

which is unramified outside the primes dividing Np and such that for all primes q | Np,

trace ρ f ,λ (Frob q) = c(q, f ) , det ρ f ,λ (Frob q) = χ (q )q k −1 . (2)

We will be concerned with trying to prove results in the opposite direction, that is to say, with
establishing criteria under which a λ -adic representation arises in this way from a modular form.
Assume
ρ 0 : Gal (Q / Q ) → GL2 (F p ) (3)

is a continuous representation with values in the algebraic closure of a finite field of characteristic p
and that det ρ 0 is odd. We say that ρ 0 is modular if ρ 0 and ρ f ,λ mod λ are isomorphic over F p
for some f and λ and some embedding of Ο f / λ in F p . Serre has conjectured that every
irreducible ρ 0 of odd determinant is modular.
If Ο is the ring of integers of a local field (containing Q p ) we will say that

ρ : Gal (Q / Q ) → GL2 (Ο ) (4)

is a lifting of ρ 0 if, for a specified embedding of the residue field of Ο in F p , ρ and ρ 0 are
isomorphic over F p . We will restrict our attention to two cases:

(I) ρ 0 is ordinary (at p) by which we mean that there is a one-dimensional subspace of Fp2 , stable
under a decomposition group at p and such that the action on the quotient space is unramified
and distinct from the action on the subspace.

(II) ρ 0 is flat (at p), meaning that as a representation of a decomposition group at p, ρ 0 is

equivalent to one that arises from a finite flat group scheme over Z p , and det ρ 0 restricted to an
inertia group at p is the cyclotomic character.

CONJECTURE.

4
Suppose that ρ : Gal (Q / Q ) → GL2 (Ο ) is an irreducible lifting of ρ 0 and that ρ is unramified
outside of a finite set of primes. There are two cases:

(i) Assume that ρ 0 is ordinary. Then if ρ is ordinary and det ρ = ε k −1χ for some integer
k ≥ 2 and some χ of finite order, ρ comes from a modular form.
(ii) Assume that ρ 0 is flat and that p is odd. Then if ρ restricted to a decomposition group
at p is equivalent to a representation on a p-divisible group, again ρ comes from a
modular form.

Now we will assume that p is an odd prime, we have the following theorem:

THEOREM 2.

Suppose that ρ 0 is irreducible and satisfies either (I) or (II) above. Suppose also that

 p −1 
(i) ρ 0 is absolutely irreducible when restricted to Q (− 1) 2 p  .
 
(ii) If q ≡ −1 mod p is ramified in ρ 0 then either ρ 0 Dq is reducible over the algebraic closure
where Dq is a decomposition group at q or ρ 0 I q is absolutely irreducible where I q is an
inertia group at q.

Then any representation ρ as in the conjecture does indeed come from a modular form.

The only condition which really seems essential to our method is the requirement that ρ 0 is
modular. The most interesting case at the moment is when p = 3 and ρ 0 can be defined over F3 .
Then since PGL2 (F3 ) ≅ S 4 every such representation is modular by the theorem of Langlands and
Tunnell. In particular, every representation into GL2 (Z 3 ) whose reduction satisfies the given
conditions is modular. We deduce:

THEOREM 3.

Suppose that E is an elliptic curve defined over Q and that ρ 0 is the Galois action on the 3-division
points. Suppose that E has the following properties:

(i) E has good or multiplicative reduction at 3.

(ii) (
ρ 0 is absolutely irreducible when restricted to Q − 3 . )
(iii) For any q ≡ −1 mod 3 either ρ 0 Dq is reducible over the algebraic closure or ρ 0 I q is
absolutely irreducible.

Then E should be modular.

The important class of semistable curves, i.e., those with square-free conductor, satisfies (i) and (iii)
but not necessarily (ii).

THEOREM 4.

5
Suppose that E is a semistable elliptic curve defined over Q. Then E is modular.

In 1986, Serre conjectured and Ribet proved a property of the Galois representation associated to
modular forms which enabled Ribet to show that Theorem 4 implies “Fermat’s Last Theorem”.
Furthermore, we have the following theorems:

THEOREM 5.

Suppose that u p + v p + w p = 0 with u , v, w ∈ Q and p ≥ 3 then uvw = 0 . (Equivalently – there are

no non-zero integers a,b,c,n with n > 2 such that a n + b n = c n .)

THEOREM 6.

Suppose that ρ 0 is irreducible and satisfies the hypothesis of the conjecture, including (I) above.
Suppose further that

(i) ρ 0 = Ind LQκ 0 for a character κ 0 of an imaginary quadratic extension L of Q which is unramified
at p.
(ii) det ρ 0 I p = ω .

Then a representation ρ as in the conjecture does indeed come from a modular form.

Wiles has worked on the Iwasawa conjecture for totally real fields and some applications of it, with
the assumption that the reduction of a given l -adic representation was reducible and tried to prove
under this hypothesis that the representation itself would have to be modular. Thence, we write p for
l because of the connections with Iwasawa theory.
In the solution to the Iwasawa conjecture for totally real fields, Wiles has introduced a new
technique in order to deal with the trivial zeroes.
It involved replacing the standard Iwasawa theory method of considering the fields in the
cyclotomic Z p -extension by a similar analysis based on a choice of infinitely many distinct primes
qi ≡ 1 mod p ni with ni → ∞ as i → ∞ . Wiles has developed further the idea of using auxiliary
primes to replace the change of field that is used in Iwasawa theory.

Let p be an odd prime. Let Σ be a finite set of primes including p and let QΣ be the maximal
extension of Q unramified outside this set and ∞ . Throughout we fix an embedding of Q , and so
also of QΣ , in C. We will also fix a choice of decomposition group Dq for all primes q in Z.
Suppose that k is a finite field characteristic p and that

ρ 0 : Gal (QΣ / Q ) → GL2 (k ) (5)

is an irreducible representation. We will assume that ρ 0 comes with its field of definition k and that
det ρ 0 is odd.
We will restrict our choice of ρ 0 further by assuming that either:

(i) ρ 0 is ordinary. The restriction of ρ 0 to the decomposition group D p has (for a suitable choice of

6
 basis) the form

χ ∗
ρ 0 D p ≈  1  (6)
 0 χ2 

where χ1 and χ 2 are homomorphisms from D p to k ∗ with χ 2 unramified. Moreover we require

that χ1 ≠ χ 2 .

(ii) ρ 0 is flat at p but not ordinary. Then ρ 0 D p is the representation associated to a finite flat group
scheme over Z p but is not ordinary in the sense of (i). We will assume also that det ρ 0 I p = ω
where I p is an inertia group at p and ω is the Teichmuller character giving the action on p th
roots of unity.

Furthermore, we have the following restrictions on the deformations:

(i) (a) Selmer deformations. In this case we assume that ρ 0 is ordinary, with notion as above, and
that the deformation has a representative ρ : Gal (QΣ / Q ) → GL2 ( A) with the property that
(for a suitable choice of basis)
 χ~ ∗ 
ρ D p ≈  1 ~ 
 0 χ2 

with χ~2 unramified, χ~ ≡ χ 2 mod m , and det ρ I p = εω −1χ1χ 2 where ε is the cyclotomic
character, ε : Gal (QΣ / Q ) → Z ∗p , giving the action on all p-power roots of unity, ω is of
order prime to p satisfying ω ≡ ε mod p , and χ1 and χ 2 are the characters of (i) viewed as
taking values in k ∗ a A∗ .

(i) (b) Ordinary deformations. The same as in (i) (a) but with no condition on the determinant.

(i) (c) Strict deformations. This is a variant on (i) (a) which we only use when ρ 0 D p is not
semisimple and not flat. We also assume that χ1χ 2−1 = ω in this case. Then a strict
deformation is an in (i) (a) except that we assume in addition that (χ~ / χ~ ) D = ε .
1 2 p

(ii) Flat (at p) deformations. We assume that each deformations ρ to GL2 ( A) has the property
that for any quotient A / a of finite order ρ D p mod a is the Galois representation associated
to the Q p -points of a finite flat group scheme over Z p .

In each of these four cases, as well as in the unrestricted case one can verify that Mazur’s use of
Schlessinger’s criteria proves the existence of a universal deformation

ρ : Gal (QΣ / Q ) → GL2 (R ) (7)

With regard the primes q ≠ p which are ramified in ρ 0 , we distinguish three special cases:

7
 χ ∗
(A) ρ 0 Dq =  1  for a suitable choice of basis, with χ1 and χ 2 unramified, χ1χ 2−1 = ω and
 χ2 
the fixed space of I q of dimension 1,
 χ 0
(B) ρ 0 I q =  q , χ q ≠ 1 , for a suitable choice of basis,
 0 1
(C) H 1 (Qq ,Wλ ) = 0 where Wλ = { f ∈ Homk (U λ ,U λ ) : tracef = 0} ≅ (Sym 2 ⊗ det −1 )ρ 0 .

Then in each case we can define a suitable deformation theory by imposing additional restrictions
on those we have already considered, namely:

ψ ∗
(A) ρ Dq =  1  for a suitable choice of basis of A2 with ψ 1 and ψ 2 unramified and
 ψ2
ψ 1ψ 2−1 = ε ;
χ 0
(B) ρ I q =  q  for a suitable choice of basis ( χ q of order prime to p, so the same character as
0 1 
above);

(C) det ρ I q = det ρ 0 I q , i.e., of order prime to p.

Thus if Μ is a set of primes in Σ distinct from p and each satisfying one of (A), (B) or (C) for ρ 0 ,
we will impose the corresponding restriction at each prime in Μ .
Thus to each set of data D = {⋅, Σ, Ο, Μ} where . is Se, str, ord, flat or unrestricted, we can associate
a deformation theory to ρ 0 provided

ρ 0 : Gal (QΣ / Q ) → GL2 (k ) (8)

is itself of type D and Ο is the ring of integers of a totally ramified extension of W (k ) ; ρ 0 is

ordinary if . is Se or ord, strict if . is strict and flat if . is flat; ρ 0 is of type Μ , i.e., of type (A), (B)
or (C) at each ramified primes q ≠ p , q ∈ Μ .

Suppose that q is a prime not dividing N. Let Γ1 (N , q ) = Γ1 ( N ) I Γ0 (q ) and let

X 1 ( N , q ) = X 1 ( N , q )/ Q be the corresponding curve. The two natural maps X 1 (N , q ) → X 1 ( N )
induced by the maps z → z and z → qz on the upper half plane permit us to define a map
J1 ( N ) × J1 ( N ) → J1 ( N , q ) . Using a theorem of Ihara, Ribet shows that this map is injective. Thus
we can define ϕ by
ϕ
0 → J 1 ( N ) × J1 ( N ) → J1 ( N , q ) . (9)

Dualizing, we define B by
ψ ϕˆ
0 → B → J1 ( N , q ) → J1 ( N ) × J 1 ( N ) → 0 .

8
Let T1 ( N , q ) be the ring of endomorphism of J1 ( N , q ) generated by the standard Hecke operators.
One can check that U p preserves B either by an explicit calculation or by noting that B is the
maximal abelian subvariety of J1 ( N , q ) with multiplicative reduction at q. We set
J 2 = J1 ( N ) × J1 ( N ) . More generally, one can consider J H ( N ) and J H (N , q ) in place of J1 ( N ) and
J1 ( N , q ) (where J H (N , q ) corresponds to X 1 ( N , q ) / H ) and we write TH ( N ) and TH ( N , q ) for the
associated Hecke rings.
(
In the following lemma if m is a maximal ideal of Τ1 Nq r −1 or Τ1 Nq r we use m(q ) to denote the ) ( )
(q )
maximal ideal of Τ1 Nq , q ( r r +1
)
compatible with m , the ring Τ1 (q )
(Nq , q ) ⊂ Τ (Nq , q )
r r +1
1
r r +1
being
the sub-ring obtained by omitting U q from the list of generators.

LEMMA 1.

If q ≠ p is a prime and r ≥ 1 then the sequence of abelian varieties

ξ1 ξ2
( ) ( ) (
0 → J1 Nq r −1 → J1 Nq r × J1 Nq r → J1 Nq r , q r +1 ) ( ) (10)

(
where ξ1 = (π 1, r o π ) ,−(π 2, r o π )
∗ ∗
) (
and ξ 2 = π 4∗, r , π 3∗, r ) induces a corresponding sequence of p-
divisible groups which becomes exact when localized at any m(q ) for which ρ m is irreducible.

Now, we have the following theorem:

THEOREM 7.

 p −1 
Assume that ρ 0 is modular and absolutely irreducible when restricted to Q (− 1) 2 p  . Assume
 
also that ρ 0 is of type (A), (B) or (C) at each q ≠ p in Σ . Then the map ϕ D : RD → ΤD (remember
that ϕ D is an isomorphism) is an isomorphism for all D associated to ρ 0 , i.e., where
D = (⋅, Σ, Ο, Μ ) with ⋅ = Se, str, fl or ord. In particular if ⋅ = Se, str or fl and f is any newform for
which ρ f , λ is a deformation of ρ 0 of type D then

# H D1 (QΣ / Q,V f ) =# (Ο /η D , f ) < ∞ (11)

where η D, f is the invariant defined in the following equation (η ) = (η D , f ) = (πˆ (1)) .

We assume that

ρ = Ind LQκ : Gal (Q / Q ) → GL2 (Ο ) (12)

is the p-adic representation associated to a character κ : Gal (L / L ) → Ο× of an imaginary

quadratic field L .
Let M ∞ be the maximal abelian p-extension of L(ν ) unramified outside p .

PROPOSITION 1.

9
There is an isomorphism

≈
1
H unr ( )
QΣ / Q, Y ∗ → Hom(Gal (M ∞ / L(ν )), (K / Ο )(ν ))
Gal ( L (ν ) / L )
(13)

1
where H unr denotes the subgroup of classes which are Selmer at p and unramified everywhere else.

Now we write 1
H str (
QΣ / Q, Yn∗ ) (where Yn∗ = Yλ∗n and similarly for Yn ) for the subgroup of
1
H unr( ) {
QΣ / Q, Yn∗ = α ∈ H unr
1
( )
QΣ / Q, Yn∗ : α p = 0inH 1 Q p , Yn∗ / Yn∗ ( ( ) )} 0
where Yn∗ ( )0
is the first step in the
filtration under D p , thus equal to Yn / Yn0 ( )∗
or equivalently to Y ∗ ( ) 0
λn where Y ∗( ) 0
is the divisible
submodule of Y ∗ on which the action of I p is via ε 2 . It follows from an examination of the action
I p on Yλ that

1
H str (QΣ / Q, Yn ) = H unr
1
(QΣ / Q, Yn ) . (14)

In the case of Y ∗ we will use the inequality

1
# H str (
QΣ / Q, Y ∗ ≤ # H unr
1
)
QΣ / Q, Y ∗ . ( ) (15)

Furthermore, for n sufficiently large the map

1
H str (
QΣ / Q, Yn∗ → H str
1
)
QΣ / Q, Y ∗ ( ) (16)

is injective.
The above map is then injective whenever the connecting homomorphism

( )
H 0 L p ∗ , (K / Ο )(ν ) → H 1 L p ∗ , (K / Ο )(ν )λn ( )
is injective, which holds for sufficiently large n. Furthermore, we have

1
# H str (QΣ / Q, Yn ) =# H 0 Q , Y 0 ( ( ) ) ## HH ((QQ,,YY )) .
∗
0
n
(17)
(
# H str QΣ / Q, Yn∗
1 p
)n 0
n
∗

Thence, setting t = inf q # (Ο / (1 − ν (q ))) if ν mod λ = 1 or t = 1 if ν mod λ ≠ 1 (17b), we get

1
# H Se (QΣ / Q, Y ) ≤ 1 ⋅ ∏ l q ⋅# Hom(Gal (M ∞ / L(ν )), (K / Ο)(ν ))Gal (L (ν ) / L ) (18)
t ∈Σ

( )
where l q = # H 0 Qq , Y ∗ for q ≠ p , l p = lim # H 0 Q p , Yn0
n →∞
( ( ) ).
∗
This follows from Proposition 1, (14)-
(17) and the elementary estimate

(
1
# H Se (QΣ / Q, Y ) / H unr
1
(QΣ / Q, Y ) ≤ ) ∏l q , (19)
{ }
q∈Σ − p

10
 ( )
which follows from the fact that # H 1 Qqunr , Y ( )Gal Q qunr / Q q
= l q . (Remember that l is the l -adic
representation).
Let w f denote the number of roots of unity ζ of L such that ζ ≡ 1 mod f ( f an integral ideal of
Ο L ). We choose an f prime to p such that w f = 1 . Then there is a grossencharacter ϕ of L
satisfying ϕ ((α )) = α for α ≡ 1 mod f . According to Weil, after fixing an embedding Q a Q p we
can associate a p-adic character ϕ p to ϕ . We choose an embedding corresponding to a prime above
p and then we find ϕ p = κ ⋅ χ for some χ of finite order and conductor prime to p.
The grossencharacter ϕ (or more precisely ϕ o N F / L ) is associated to a (unique) elliptic curve E
defined over F = L( f ) , the ray class field of conductor f , with complex multiplication by Ο L and
isomorphic over C to C / Ο L . We may even fix a Weierstrass model of E over Ο F which has good
reduction at all primes above p . For each prime Β of F above p we have a formal group ÊΒ ,
and this is a relative Lubin-Tate group with respect to FΒ over L p . We let λ = λÊ be the
Β

logarithm of this formal group.

Let U ∞ be the product of the principal local units at the primes above p of L fp ∞ ; i.e., ( )
U ∞ = ∏U ∞ ,Β where U ∞ , Β = limU n , Β .
←
Β p

To an element u = lim un ∈ U ∞ we can associate a power series f u , Β (Τ ) ∈ ΟΒ [Τ] where ΟΒ is the

×
←

ring of integers of FΒ . For Β we will choose the prime above p corresponding to our chosen
embedding Q a Q p . This power series satisfies un, Β = ( f u , Β )(ωn ) for all n > 0, n ≡ 0(d ) where
[
d = FΒ : L p ] and {ωn } is chosen as an inverse system of π n division points of ÊΒ . We define a
homomorphism δ k : U ∞ → ΟΒ by

k
 1 d 
δ k (u ) := δ k , Β (u ) =  log f u , Β (Τ) Τ = 0 . (20)
 λ 'Eˆ (Τ) dΤ 
 Β 

Then
δ k (uτ ) = θ (τ )k δ k (u ) (21) for τ ∈ Gal (F / F )

[ ]
where θ denotes the action on E p ∞ . Now θ = ϕ p on Gal (F / F ) . We want a homomorphism
on u∞ with a transformation property corresponding to ν on all of Gal (L / L ) . We observe that
ν = ϕ p2 on Gal (F / F ) .
Let S be a set of coset representatives for Gal (L / L ) / Gal (L / F ) and define

( )
Φ 2 (u ) = ∑ν −1 (σ )δ 2 u σ ∈ ΟΒ [ν ] . (22)
σ ∈S

Each term is independent of the choice of coset representative by (17b) and it is easily checked that

11
 ( )
Φ 2 u σ = ν (σ )Φ 2 (u ) .

It takes integral values in ΟΒ [ν ] . Let U ∞ (ν ) denote the product of the groups of local principal units
at the primes above p of the field L(ν ) . Then Φ 2 factors through U ∞ (ν ) and thus defines a
continuous homomorphism
Φ 2 : U ∞ (ν ) → C p .

Let C∞ be the group of projective limits of elliptic units in L(ν ) . Then we have a crucial theorem of
Rubin:

THEOREM 8.

There is an equality of characteristic ideals as Λ = Z p [[Gal (L(ν ) / L )]] -modules:

char ∧ (Gal (M ∞ / L(ν ))) = char ∧ (U ∞ (ν ) / C∞ ) .

Let ν 0 = ν mod λ . For any Z p [Gal (L(ν 0 ) / L )]-module X we write X (ν 0 ) for the maximal quotient
of X ⊗ Ο on which the action of Gal (L(ν 0 ) / L ) is via the Teichmuller lift of ν 0 . Since
Zp

Gal (L(ν ) / L ) decomposes into a direct product of a pro-p group and a group of order prime to p,

Gal (L(ν ) / L ) ≅ Gal (L(ν ) / L(ν 0 )) × Gal (L(ν 0 ) / L ) ,

we can also consider any Z p [[Gal (L(ν ) / L )]] -module also as a Z p [Gal (L(ν 0 ) / L )]-module. In
(ν 0 )
particular X (ν 0 ) is a module over Z p [Gal (L(ν 0 ) / L )] ≅ Ο . Also Λ(ν 0 ) ≅ Ο[[Τ]] .
(ν 0 )
Now according to results of Iwasawa, U ∞ (ν ) is a free Λ(ν 0 ) -module of rank one. We extend Φ 2
(ν 0 )
Ο -linearly to U ∞ (ν ) ⊗ Z p Ο and it then factors through U ∞ (ν ) . Suppose that u is a generator of
(ν )
U ∞ (ν ) 0 and β an element of C∞(ν 0 ) . Then f (γ − 1)u = β for some f (Τ) ∈ Ο[[Τ]] and γ a
topological generator of Gal (L(ν ) / L(ν 0 )) . Computing Φ 2 on both u and β gives

f (ν (γ ) − 1) = φ2 (β ) / Φ 2 (u ) . (23)

We have that ν can be interpreted as the grossencharacter whose associated p-adic character , via
the chosen embedding Q a Q p , is ν , and ν is the complex conjugate of ν .
Furthermore, we can compute Φ 2 (u ) by choosing a special local unit and showing that Φ 2 (u ) is a
p-adic unit.
Now, if we have that
1
# H Se (
(QΣ / Q, Y ) ≤# Ο / Ω − 2 L f 0 (2,ν ) ⋅ ∏ l q , )
q∈Σ

and # (Ο / hL ) ⋅ ∏l q , (24)
q∈Σ − { p }

12
 (
where l q = # H 0 Qq , ((K / Ο )(ψ ) ⊕ K / Ο )
∗
) and hL is the class number of Ο L , combining these we
obtain the following relation:

1
# H Se ( )
(QΣ / Q,V ) ≤# Ο / Ω − 2 L f 0 (2,ν ) ⋅# (Ο / hL ) ⋅ ∏ l q , (25)
q∈Σ

( ) ( ( )) ∗
where l q = # H 0 Qq ,V ∗ (for q ≠ p ), l p = # H 0 Q p , Y 0 . (Also here, we remember that l is p-
adic).
Let ρ 0 be an irreducible representation as in (5). Suppose that f is a newform of weight 2 and
level N, λ a prime of Ο f above p and ρ f , λ a deformation of ρ 0 . Let m be the kernel of the
homomorphism Τ1 ( N ) → Ο f / λ arising from f .
We now give an explicit formula for η developed by Hida by interpreting , in terms of the cup
product pairing on the cohomology of X 1 ( N ) , and then in terms of the Petersson inner product of
f with itself. Let
(, ) : H 1 (X 1 (N ), Ο f )× H 1 (X 1 (N ), Ο f ) → Ο f (26)

be the cup product pairing with Ο f as coefficients. Let p f be the minimal prime of Τ1 ( N ) ⊗ O f
associated to f , and let
L f = H 1 (X 1 ( N ), Ο f ) p f . [ ]
If f = ∑ an q n let f ρ = ∑ an q n . Then f ρ is again a newform and we define L f ρ by replacing
f by f ρ in the definition of L f . Then the pairing (,) induces another by restriction

(, ) : L f × L f ρ → Οf . (27)

Replacing Ο by the localization of Ο f at p (if necessary) we can assume that L f and L f ρ are free
of rank 2 and direct summands as Ο f -modules of the respective cohomology groups. Let δ1 , δ 2 be
a basis of L f . Then also δ1 , δ 2 is a basis of L f ρ = L f . Here complex conjugation acts on
H 1 (X 1 ( N ), Ο f ) via its action on Ο f . We can then verify that

(δ ,δ ) := det (δ ,δ ) i j

is an element of Ο f whose image in Ο f , λ is given by π η 2 (unit). ( )

To give a more useful expression for (δ , δ ) we observe that f and f ρ can be viewed as elements
of H 1 ( X 1 ( N ), C ) ≅ H 1DR ( X 1 (N ), C ) via f a f (z )dz , { }
f ρ a f ρ dz . Then f , f ρ form a basis for
L f ⊗Ο f C . Similarly {f , f } form
ρ
a basis for L f ρ ⊗Ο f C . ( )
Define the vectors ω1 = f , f ρ ,
ω2 = ( f , f ρ ) and write ω1 = Cδ and ω2 = C δ with C ∈ M 2 (C ) . Then writing f1 = f , f 2 = f ρ
we set
(ω , ω ) := det (( fi , f j )) = (δ , δ )det (CC ).

13
Now (ω , ω ) is given explicitly in terms of the (non-normalized) Petersson inner product , :
(ω , ω ) = −4 f, f =∫
2
f, f where ffdxdy . Hence, we have the following equation:
ℑ / Γ1 ( N )

2
(ω , ω ) = −4 ∫ℑ / Γ ( N ) ffdxdy  . (28)
 1 

To compute det (C ) we consider integrals over classes in H1 (X 1 ( N ), Ο f ) . By Poincaré duality there

exist classes c1 ,c2 in H1 (X 1 ( N ), Ο f ) such that det ∫ δ i  is a unit in Ο f . Hence det C generates
 cj 
 
the same Ο f -module as is generated by det ∫ f i  for all such choices of classes ( c1 ,c2 ) and
  j 
c

 
with { f1 , f 2 } = f , f ρ . Letting u f be a generator of the Ο f -module det ∫ f i  we have the
{ }
  c j 
following formula of Hida:
π (η 2 ) = f , f
2
/ u f u f × (unit in Ο f , λ ).

Now, we choose a (primitive) grossencharacter ϕ on L together with an embedding Q a Q p

corresponding to the prime p above p such that the induced p-adic character ϕ p has the properties:
(i) ϕ p mod p = κ 0 ( p = maximal ideal of Q p ).
(ii) ϕ p factors through an abelian extension isomorphic to Z p ⊕ T with T of finite order prime to
p.
(iii) ϕ ((α )) = α for α ≡ 1( f ) for some integral ideal f prime to p.

Let p0 = kerψ f : Τ1 ( N ) → Ο f and let A f = J1 ( N ) / p0 J1 ( N ) be the abelian variety associated to f by

Shimura. Over F + there is an isogeny Af / F + ≈ E/ F + ( )
d
[
where d = Ο f : Ζ . ]
We have that the p-adic Galois representation associated to the Tate modules on each side are
( )
equivalent to Ind FF ϕ0 ⊗ Ζ p K f , p where K f , p = Ο f ⊗ Q p and where ϕ p : Gal (F / F ) → Ζ×p is the p-
+

adic character associated to ϕ and restricted to F .We now give an expression for fϕ , fϕ in terms
of the L-function of ϕ . We note that LN (2,ν ) = LN (2,ν ) = LN 2,ϕ 2 χˆ and remember that ν is the p- ( )
adic character, and ν is the complex conjugate of ν , we have that:
 
1 2  1 
fϕ , fϕ =
16π 3
N ∏ 1 −  LN 2,ϕ 2 χˆ LN (1,ψ ) , (29) ( )
 qq∉SN  q 
 ϕ 

where χ is the character of fϕ and χ̂ its restriction to L ; ψ is the quadratic character

associated to L ; LN ( ) denotes that the Euler factors for primes dividing N have been removed;
Sϕ is the set of primes q N such that q = qq ' with q | cond ϕ and q, q ' primes of L , not
necessarily distinct.

THEOREM 9.
14
Suppose that ρ 0 as in (5) is an irreducible representation of odd determinant such that
ρ 0 = Ind LQκ 0 for a character κ 0 of an imaginary quadratic extension L of Q which is unramified
at p. Assume also that:
(i) det ρ 0 Ip =ω ;
(ii) ρ 0 is ordinary.
Then for every D = (⋅, Σ, Ο, φ ) such that ρ 0 is of type D with ⋅ = Se or ord,

RD ≅ ΤD

and ΤD is a complete intersection.

COROLLARY.

For any ρ 0 as in the theorem suppose that

ρ : Gal (Q / Q ) → GL2 (Ο )

is a continuous representation with values in the ring of integers of a local field, unramified outside
a finite set of primes, satisfying ρ ≅ ρ 0 when viewed as representations to GL2 (Fp ) . Suppose
further that:
(i) ρ Dp is ordinary;

(ii) det ρ Ip = χε k −1 with χ of finite order, k ≥ 2 .

Then ρ is associated to a modular form of weight k .

THEOREM 10. (Langlands-Tunnell)

Suppose that ρ : Gal (Q / Q ) → GL2 (C ) is a continuous irreducible representation whose image is

finite and solvable. Suppose further that det ρ is odd. Then there exists a weight one newform f
such that L(s, f ) = L(s, ρ ) up to finitely many Euler factors.

Suppose then that

ρ 0 : Gal (Q / Q ) → GL2 (F3 )
is an irreducible representation of odd determinant. This representation is modular in the sense that
over F3 , ρ 0 ≈ ρ g , µ mod µ for some pair ( g , µ ) with g some newform of weight 2. There exists a
representation
([ ])
i : GL2 (F3 ) a GL2 Ζ − 2 ⊂ GL2 (C ).

By composing i with an automorphism of GL2 (F3 ) if necessary we can assume that i induces the
( )
identity on reduction mod 1 + − 2 . So if we consider i o ρ 0 : Gal (Q / Q ) → GL2 (C ) we obtain an
irreducible representation which is easily seen to be odd and whose image is solvable.
Now pick a modular form E of weight one such that E ≡ 1(3) . For example, we can take E = 6E1, χ
where E1, χ is the Eisenstein series with Mellin transform given by ζ (s )ζ (s, χ ) for χ the quadratic

15
 ( )
character associated to Q − 3 . Then fE ≡ f mod 3 and using the Deligne-Serre lemma we can
find an eigenform g ' of weight 2 with the same eigenvalues as f modulo a prime µ ' above
(1 + )
− 2 . There is a newform g of weight 2 which has the same eigenvalues as g ' for almost all
( )
Tl ’s, and we replace ( g ' , µ ') by ( g , µ ) for some prime µ above 1 + − 2 . Then the pair ( g , µ )
satisfies our requirements for a suitable choice of µ (compatible with µ ' ).
We can apply this to an elliptic curve E defined over Q , and we have the following fundamental
theorems:

THEOREM 11.

All semistable elliptic curves over Q are modular.

THEOREM 12.

Suppose that E is an elliptic curve defined over Q with the following properties:
(i) E has good or multiplicative reduction at 3, 5,
(ii) For p = 3, 5 and for any prime q ≡ −1 mod p either ρ E , p Dq is reducible over Fp or ρ E , p I q is
irreducible over Fp .
Then E is modular.

Chapter 2.

Further mathematical aspects concerning the Fermat’s Last Theorem

2.1 On the modular forms, Euler products, Shimura map and automorphic L-functions.

A. Modular forms[2]

We know that there is a direct relation with elliptic curves, via the concept of modularity of elliptic
curves over Q .
Let E be an elliptic curve over Q , given by some Weierstrass equation. Such a Weierstrass
equation can be chosen to have its coefficients in Z . A Weierstrass equation for E with
coefficients in Z is called minimal if its discriminant is minimal among all Weierstrass equations
for E with coefficients in Z ; this discriminant then only depends on E and will be denoted
discr( E ). Thence, E has a Weierstrass minimal model over Z , that will be denoted by EZ .
For each prime number p, we let EF p denote the curve over Fp given by reducing a minimal
Weierstrass equation modulo p; it is the fibre of EZ over Fp . The curve EF p is smooth if and only if
p does not divide discr( E ).
The possible singular fibres have exactly one singular point: an ordinary double point with rational
tangents, or with conjugate tangents, or an ordinary cusp. The three types of reduction are called
16
split multiplicative, non-split multiplicative and additive, respectively, after the type of group law
that one gets on the complement of the singular point. For each p we then get an integer a p by
requiring the following identity:
p + 1 − a p = # E (Fp ) . (1)

This means that for all p, a p is the trace of Fp on the degree one étale cohomology of EF p , with
coefficients in Fl , or in Z / l n Z or in the l -adic numbers Z l . For p not dividing discr( E ) we know
that a p ≤ 2 p1 / 2 . If EF p is multiplicative, then a p = 1 or – 1 in the split and non-split case. If EF p
is additive, then a p = 0 . We also define, for each p an element ε ( p ) in {0,1} by setting ε ( p ) = 1 for
p not dividing discr( E ). The Hasse-Weil L-function of E is then defined as:

LE (s ) = ∏ LE , p (s ) , (
LE , p (s ) = 1 − a p p − s + ε ( p ) pp − 2 s )−1
, (2)
p

for s in C with R(s ) > 3 / 2 . We note that for all p and for all l ≠ p we have the identity:

( ( ))
1 − a pt + ε ( p )t 2 = det 1 − tFp∗ , H 1 EF , et , Ql . (3)

We use étale cohomology with coefficients in Ql , the field of l -adic numbers, and not in Fl .
The function LE was conjectured to have a holomorphic continuation over all of C , and to satisfy a
certain precisely given functional equation relating the values at s and 2 − s . In that functional
equation appears a certain positive integer N E called the conductor of E , composed of the primes p
dividing discr( E ) with exponents that depend on the behaviour of E at p, i.e., on EZ p . This
conjecture on continuation and functional equation was proved for semistable E (i.e., E such that
there is no p where E has additive reduction) by Wiles and Taylor-Wiles, and in the general case
by Breuil, Conrad, Diamond and Taylor. In fact, the continuation and functional equation are direct
consequences of the modularity of E that was proved by Wiles, Taylor-Wiles, etc.
The weak Birch and Swinnerton-Dyer conjecture says that the dimension of the Q -vector space
Q ⊗ E (Q ) is equal to the order of vanishing of LE at 1. Anyway, the function LE gives us integers
an for all n ≥ 1 as follows:
LE (s ) = ∑ an n − s , for R(s ) > 3 / 2 . (4)
n ≥1

From these an one can then consider the following function:

f E : H = {τ ∈ C ℑ(τ ) > 0} → C , τ a ∑ an e 2πinτ . (5)

n ≥1

Equivalently, we have:
f E = ∑ an q n , with q : H → C , τ a e 2πiτ . (6)
n ≥1

A more conceptual way to state the relation between LE and f E is to say that LE is obtained, up to
elementary factors, as the Mellin transform of f E :

17
 ∞ dt
∫ f E (it )t s = (2π ) Γ(s )LE (s ) , for R(s ) > 3 / 2 .
−s
(7)
0 t

Hence, we can finally state what the modularity of E means:

f E is a modular form of weight two for the congruence subgroup Γ0 ( N E ) of SL2 (Z ) .
The last statement means that f E has an enormous amount of symmetry.
A typical example of a modular form of weight higher than two is the discriminant modular form,
usually denoted ∆ . One way to view ∆ is as the holomorphic function on the upper half plane H
given by:
∆ = q∏ 1 − q n ( )
24
, (8)
n ≥1

where q is the function from H to C given by z a exp(2πiz ) . The coefficients in the power series
expansion:
∆ = ∑τ (n )q n (9)
n ≥1

define the famous Ramanujan τ -function.

To say that ∆ is a modular form of weight 12 for the group SL2 (Z ) means that for all elements
a b
  of SL2 (Z ) the following identity holds for all z in H :
c d 

 az + b 
 = (cz + d ) ∆( z ) ,
12
∆ (10)
 cz + d 

which is equivalent to saying that the multi-differential form ∆( z )(dz ) is invariant under the
⊗6

 1 1  0 − 1
action of SL2 (Z ) . As SL2 (Z ) is generated by the elements   and   , it suffices to
 0 1 1 0 
check the identity in (10) for these two elements. The fact that ∆ is q times a power series in q
means that ∆ is a cusp form: it vanishes at “ q = 0 ”. It is a fact that ∆ is the first example of a non-
zero cusp form for SL2 (Z ) : there is no non-zero cusp form for SL2 (Z ) of weight smaller than 12,
i.e., there are no non-zero holomorphic functions on H satisfying (10) with the exponent 12
replaced by a smaller integer, whose Laurent series expansion in q is q times a power series.
Moreover, the C -vector space of such functions of weight 12 is one-dimensional, and hence ∆ is a
basis of it.
The one-dimensionality of this space has as a consequence that ∆ is an eigenform for certain
operators on this space, called Hecke operators, that arise from the action on H of GL2 (Q ) , the
+

subgroup of GL2 (Q ) of elements whose determinant is positive. This fact explains that the
coefficients τ (n ) satisfy certain relations which are summarised by the following identity of
Dirichlet series:
(
L∆ (s ) := ∑τ (n )n − s = ∏ 1 − τ ( p ) p − s + p11 p − 2 s )−1
. (11)
n ≥1 p

These relations:
τ (mn ) = τ (m )τ (n ) if m and n are relatively prime;

18
 τ ( p n ) = τ ( p n −1 )τ ( p ) − p11τ ( p n − 2 ) if p is prime and n ≥ 2

were conjectured by Ramanujan, and proved by Mordell. Using these identities, τ (n ) can be
expressed in terms of the τ ( p ) for p dividing n . As L∆ is the Mellin transform of ∆ , L∆ is
holomorphic on C , and satisfies the functional equation (Hecke):

(2π )− (12 − s ) Γ(12 − s )L∆ (12 − s ) = (2π )− s Γ(s )L∆ (s ) . (12)

The famous Ramanujan conjecture states that for all primes p one has the inequality:

τ ( p ) < 2 p11 / 2 , (13)

or, equivalently, that the complex roots of the polynomial x 2 − τ ( p )x + p11 are complex conjugates
of each other, and hence are of absolute value p11 / 2 .

B. Euler products[3]

We know that the infinite series

∞
1
∑n
n =1
s
, (14)

converges for R(s ) > 1 and gives rise by analytic continuation to a meromorphic function ζ (s ) in
C . For R(s ) > 1 ζ (s ) admits the absolutely convergent infinite product expansion

1
∏1− p
p
−s
, (15)

taken over the set of primes. This “Euler product” may be regarded as an analytic formulation of the
principle of unique factorization in the ring Z of integers. It is, as well, the product taken over all
the non-Archimedean completions of the rational field Q (which completions Q p are indexed by
the set of primes) of the “Mellin transform” in Q p

1
ξ p (s ) = , (16)
1 − p−s

(where the Mellin transform is, more or less, Fourier transform on the multiplicative group.
∞
Classically, the Mellin transform ϕ of f is given formally by ϕ (s ) = ∫ f ( x )x s (dx / x ) . (17))
0

of the canonical “Gaussian density” Φ p ( x ) = 1 if x ∈ closure of Z in Q p ; 0 otherwise, which

Gaussian density is equal to its own Fourier transform. For the Archimedean completion Q∞ = R of
the rational field Q one forms the classical Mellin transform

ξ ∞ (s ) = π −( s / 2 )Γ(s / 2) (18)

of the classical Gaussian density

19
 Φ ∞ ( x ) = e −πx , (19)
2

(which also is equal to its own Fourier transform). Then the function

ξ (s ) = ξ ∞ (s )ζ (s ) = ∏ ξ p (s ) (20)
p ≤∞

is meromorphic in C , and satisfies the functional equation

ξ (1 − s ) = ξ (s ) . (21)

The connection of Riemann’s ζ -function with the subject of modular forms begins with the
observation that ζ (2s ) is essentially the Mellin transform of θ I ( x ) = θ (ix ) − 1 , where θ , which is a
modular form of weight 1/2 and level 8, is defined in the upper-half plane H by the formula

θ (τ ) = ∑ exp(πiτm ).
m∈Z
2
(22)

In fact, one of the classical proofs of the functional equation (21) is given by applying the Poisson
( )
summation formula to the function x a exp πiτx 2 , while observing that the substitution
s a (1 / 2) − s for ζ (2s ) corresponds in the upper-half plane to the substitution τ a −1 / τ for the
theta series. If f is a cuspform for a congruence group Γ containing

 1 1
T =   , (23)
 0 1

and so, consequently, f (τ + 1) = f (τ ) , then one has the following Fourier expansion

∞
f (τ ) = ∑ cme 2πimτ . (24)
m =1

The Mellin transform ϕ (s ) of f I leads to the Dirichlet series

∞
ϕ (s ) = ∑ cm m − s , (25)
m =1

which may be seen to have a positive abscissa of convergence.

For the “modular group” Γ(1) the Dirichlet series associated to every cuspform of weight w admits
an analytic continuation with functional equation under the substitution s a w − s . Since Γ(1) is
generated by the two matrices T and
 0 1
W =   (26)
 −1 0

and since the functional equation of a modular form f relative to T is reflected in the formation of
the Fourier series (24), the condition that an absolutely convergent series (24) is a modular form for
Γ(1) is the functional equation for a modular form relative solely to W .

20
Observing that the formula
dx 2 + dy 2
ds 2 = for τ = x + iy ∈ H , (27)
y2

gives a (the hyperbolic) SL2 (R ) -invariant metric in H with associated invariant measure

dxdy
dµ = , (28)
y2

one introduces the Petersson (Hermitian) inner product in the space of cuspform of weight w for Γ
with the definition:
f , g = ∫ f (τ )g (τ )ℑ(τ ) dµ (τ ) . (29) (see also page 13 eq. (28))
w
H /Γ

(Integration over the quotient H / Γ makes sense since the integrand f (τ )g (τ ) y w (30) is Γ -
invariant).
For the modular group Γ(1) the n th Hecke operator T (n ) = Tw (n ) is the linear endomorphism of the
space of cuspforms of weight w arising from the following considerations. Let S n be the set of
2 × 2 matrices in Z with determinant n . For

a b 
M =   ∈ S n (31)
c d 

and for a function f in H one defines

(M ⋅w f )(τ ) = det (M )w −1 (cτ + d )− w f (τ ) , (32)

and then, observing that Γ(1) under ⋅w acts trivially on the modular forms of weight w , one may
define the Hecke operator Tw (n ) by

Tw (n )( f ) = ∑ (M ⋅ f )(τ ) ,
M ∈S n / Γ (1)
w (33)

where the quotient S n / Γ(1) refers to the action of Γ(1) by left multiplication on the set S n . One
finds for m, n coprime that
T (mn ) = T (m )T (n ) , (34)

and furthermore one has

( ) ( )
T p e +1 = T p e T ( p ) − p w −1T p e −1 . ( ) (35)

Consequently, the operators T (n ) commute with each other, and, therefore, generate a commutative
algebra of endomorphisms of the space of cusp forms of weight w for Γ(1) . It is not difficult to see
that the Hecke operators are self-adjoint for the Petersson inner product on the space of cuspforms.
Consequently, the space of cuspforms of weight w admits a basis of simultaneous eigenforms for

21
the Hecke algebra. A “Hecke eigencuspform” is said to be normalized if its Fourier coefficient
c1 = 1 . If f is a normalized Hecke eigencuspform, then:

(i) The Fourier coefficient cm of f is the eigenvalue of f for T (m ) .

(ii) The Fourier coefficients c(m ) = cm of f satisfy
c(mn ) = c(m )c(n ) for m, n coprime, and
c( p e +1 ) = c( p e )c( p ) − p w −1c( p e −1 ) for p prime.

Consequently, the Dirichlet series associated with a simultaneous Hecke eigencuspform of level 1
and weight w admits an Euler product

1
ϕ (s ) = ∏ . (36)
p 1 − c p p + p w −1− 2 s
−s

For example, when f is the unique normalized cuspform ∆ of level 1 and weight 12, one has

1
ϕ (s ) = ∏ , (37) (in fact, if w = 12 , then w − 1 − 2 s = 11 − 2 s )
p 1 − τ ( p ) p − s + p11− 2 s

where c p = τ ( p ) is the function τ of Ramanujan.

C. Shimura map[3]

Shimura showed for a given WN -compatible Hecke eigencuspform f of weight 2 for the group
Γ0 ( N ) with rational Fourier coefficients how to construct an elliptic curve E f defined over Q such
that the Dirichlet series ϕ (s ) associated with f is the same as the L -function L(E f , s ) .
Let Γ be a congruence subgroup of SL2 (Z ) , and let X (Γ ) denote the compact Riemann surface
H ∗ / Γ . The inclusion of Γ in Γ(1) induces a “branched covering”

X (Γ ) → X (1) ≅ P1 . (38)

One may use the elementary Riemann-Hurwitz formula from combinatorial topology to determine
the Euler number, and consequently the genus, of X (Γ ) . The genus is the dimension of the space of
cuspforms of weight 2. Even when the genus is zero one obtains embeddings of X (Γ ) in projective
spaces P r through holomorphic maps

τ a ( f 0 (τ ), f1 (τ ),..., f r (τ )) , (39)

where f 0 , f1 ,..., f r is a basis of the space of modular forms of weight w with w sufficiently large.

22
Using the corresponding projective embedding one finds a model for X 0 ( N ) = X (Γ0 ( N )) over Q ,
i.e., an algebraic curve defined over Q in projective space that is isomorphic as a compact Riemann
surface to X 0 (N ) .
Associated with any “complete non-singular” algebraic curve X of genus g is a complex torus, the
“Jacobian” J ( X ) of X , that is the quotient of g -dimensional complex vector space C g by the
lattice Ω generated by the “period matrix”, which is the g × 2 g matrix in C obtained by
integrating each of the g members ωi of a basis of the space of holomorphic differentials over each
of the 2 g loops in X representing the members of a homology basis in dimension 1. Furthermore,
if one picks a base point z0 in X , then for any z in X , the path integral from z0 to z of each of
the g holomorphic differentials is well-defined modulo the periods of the differential. One obtains
a holomorphic map X → J ( X ) from the formula

z a  ∫ ω1 ,..., ∫ ω g  mod Ω .
z z
(40)
 z0 z0 

This map is universal for pointed holomorphic maps from X to complex tori. Furthermore, the
Jacobian J ( X ) is an algebraic variety that admits definition over any field of definition for X and
z0 , and the universal map also admits definition over any such field. The complex tori that admit
embeddings in projective space are the abelian group objects in the category of projective varieties.
They are called abelian varieties. Every abelian variety is isogenous to the product of “simple”
abelian varieties: abelian varieties having no abelian subvarieties. Shimura showed that one of the
simple isogeny factors of J ( X 0 ( N )) is an elliptic curve E f defined over Q characterized by the
fact that its one-dimensional space of holomorphic differentials induces on X 0 (N ) , via the
composition of the universal map with projection on E f , the one-dimensional space of differentials
on X 0 (N ) determined by the cuspform f .
He showed further that L(E f , s ) is the Dirichlet series ϕ (s ) with Euler product given by f . An
elliptic curve E defined over Q is said to be modular if it is isogenous to E f for some WN -
compatible Hecke eigencuspform of weight 2 for Γ0 ( N ) . Equivalently E is modular if and only if
L(E , s ) is the Dirichlet series given by such a cuspform. The Shimura-Taniyama-Weil Conjecture
states that every elliptic curve defined over Q is modular. Shimura showed that this conjecture is
true in the special case where the Z -module rank of the ring of endomorphisms of E is grater than
one. In this case the point τ of the upper-half plane corresponding to E (C ) is a quadratic imaginary
number, and L(E , s ) is a number-theoretic L -function associated with the corresponding imaginary
quadratic number field.

D. Automorphic L -functions[4]

Talking about zeta functions in general one inevitably is led to start with the Riemann zeta function
ζ (s ) . It is defined as a Dirichlet series:

∞
ζ (s ) = ∑ n − s , (41)
n =1

23
which converges for each complex number s of real part greater than one. In the same region it
possesses a representation as a Mellin integral:

1 ∞ 1 s dt
Γ(s ) ∫0 et − 1 t
ζ (s ) = t . (42)

Let f be a cusp form of weight 2k for some natural number k , i.e., the function f is holomorphic
on the upper half plane H in C , and has a certain invariance property under the action of the
modular group SL2 (Z ) on H . Then f admits a Fourier expansion

∞
f ( z ) = ∑ ane 2πizn . (43)
n =1

Define its L -function for Re(s ) > 1 by

∞
an
L( f , s ) = ∑ s
. (44)
n =1 n

The easily established integral representation

∞ dt
Lˆ ( f , s ) = (2π ) Γ(s )L( f , s ) = ∫ f (it )t s ,
−s
(45)
0 t

implies that L( f , s ) extends to an entire function satisfying the functional equation

Lˆ ( f , s ) = (− 1) Lˆ ( f ,2k − s ) . With Λ ( f , s ) = Lˆ ( f ,2ks ) this becomes
k

Λ ( f , s ) = (− 1) Λ ( f ,1 − s ) .
k
(46)

This construction can be extended to cusp forms for suitable subgroups of the modular group. These
L -functions look like purely analytical objects. Thus it was particularly daring of A. Weil, G.
Shimura, and Y. Taniyama in 1955 to propose the conjecture that the zeta function of any elliptic
curve over Q coincides with a Λ ( f , s ) for a suitable cusp form f . This conjecture was proved in
part by A. Wiles and R. Taylor providing a proof of Fermat’s Last Theorem as a consequence.
The upper half plane is a homogeneous space of the group SL2 (R ) , and so cusp forms may be
viewed as functions on this group, in particular, they are vectors in the natural unitary
representation of SL2 (R ) on the space

L2 (SL2 (Z ) \ SL2 (R )) . (47)

Going even further one can extend this quotient space to the quotient of the adele group GL2 ( A)
modulo its discrete subgroup GL2 (Q ) , so cusp forms become vectors in

(
L2 GL2 (Q ) \ GL2 ( A) ,
1
) (48)

24
where GL2 ( A) denotes the set of all matrices in GL2 ( A) whose determinant has absolute value one.
1

Now GL2 can be replaced by GLn for n ∈ N and one can imitate the methods of Tate’s thesis (the
case n = 1) to arrive at a much more general definition of an automorphic L -function: this is an
Euler product L(π , s ) attached to an automorphic representation π of GLn ( A) , i.e., an irreducible
1

( )
subrepresentation π of L2 GLn (Q ) \ GLn ( A) . As in the GL1 -case it has an integral representation
1

as a Mellin transform and it extends to a meromorphic representation, which is entire if π is

cuspidal and n > 1 . Furthermore it satisfies a functional equation

L(π , s ) = ε (π , s )L(π~,1 − s ) , (49)

where π~ is the contragredient representation and ε (π , s ) is a constant multiplied by an exponential.

We conclude remember that extending the Weil-Shimura-Taniyama conjecture, R.P. Langlands
conjectured in the 1960s that any motivic L -function coincides with L(π , s ) for some cuspidal π .

2.2 On some mathematical applications of the Mellin transform.[5]

Harmonic sums are sums of the form

G ( x ) = ∑ λk g (µ k x ) , (50)
k

where the λk are the amplitudes, the µ k are the frequencies and g ( x ) is the base function. We
consider harmonic sums because we wish to evaluate G ( x ) at a set of particular points x0 , x1 ,... or at
all x ∈ R .

Definition of the harmonic sum and computation of the appropriate Mellin transform.

Now, let λk = 1 / k , µ k = 1 / k and g ( x ) = x / (1 + x ) = 1 / (1 + 1 / x ) ; and we consider the harmonic sum

1 x/k 1 1 
h( x ) = ∑ λk g (µ k x ) = ∑ = ∑ − . (51)
x k 1+ x / k k x+k

This sum is of interest because

1 1  1 1 n
1
h(n ) = ∑  −  ∑
= − ∑ = ∑ = Hn , (52)
k n+k  k k = n +1 k k =1 k

the n th harmonic number.

The principal operation in the evaluation of harmonic sums is the computation of the Mellin
transform of the base function g ( y ) and the computation of the Dirichlet generating function Λ (s ) .
We first compute the transform of the base function. We have Μ[1 / (1 + x ); s ] = π / sin (πs ) and
hence
 x  π
Μ ; s = − . (53)
1 + x  sin (πs )

Now we compute the Dirichlet generating function Λ (s ) . We have

25
 1 1
Λ (s ) = ∑ k s = ∑ 1− s = ζ (1 − s ) . (54)
k k

We conclude that the Mellin transform of h( x ) is

π
− ζ (1 − s ) . (55)
sin (πs )

Inversion of the map.

Now, by Mellin inversion we obtain:

 π 
Μ −1 − ζ (1 − s ); x  = h( x ) . (56)
 sin (πs ) 

This is equivalent to the inversion integral

c + i∞ π 
∫c − i∞
 −
 sin (πs )
ζ (1 − s ) x − s ds = h( x ) .

(57)

This integral representation permits the computation of h( x ) , because the integral can evaluated by
the Cauchy Residue theorem, i.e., it is a sum of residues of h∗ (s )x − s .

Computation of the poles of the transform function and the corresponding terms in the asymptotic
expansion.

We use the fact that

h( x ) ≈ − ∑ Re s(h (s )x ∗ −s
)
; s = ς , (58)
(
ς ∈Sing h ( s ) x
∗ −s
)I H

where H is the right half-plane, chosen for an expansion at infinity. We must compute the set of
( )
poles Sing h∗ (s )x − s I H and map them back to the terms of the expansion of h( x ) . The poles of
h (s ) in the right half-plane are at s = 0 , where we have a double pole and
∗

1 γ
h ∗ (s ) = − + ... (59)
s2 s

and at s = k , k ∈ Z + , where we have

ζ (1 − k )
h∗ (s ) = −(− 1)
k
+ ... . (60)
s−k

These poles map back to − log x − γ (61) and −

1
for k = 1 , −
(− 1) Bk 1 for k ≥ 2 . (62)
k

2x k xk

We conclude that Harmonic numbers satisfy the asymptotic expansion

26
 H n ≈ log n + γ +
1
+∑
(− 1) Bk 1 .k
(63)
2n k ≥ 2 k nk

This expansion is exact; it converges for n ≥ 1 .

The Mellin transform maps the space of functions that are integrable along the positive real line to
that of complex functions that are analytic on a vertical strip of the complex plane. This strip may in
many cases be extended to a larger domain. The map is given by the fundamental formula:

+∞
Μ[ f ( x ); s ] = f ∗ (s ) = ∫ f ( x )x s −1dx . (64)
0

The Mellin-Perron formula is a specific instance of generalized Mellin summation. The traditional
proof uses the “discontinuous factor” described by the following lemma:

m
1 c + i∞ ys 1  1
φ (y) =
2πi ∫c − i∞ s (s + 1)...(s + m )
ds = 1 −  if 1 ≤ y ;
m!  y

1 c + i∞ ys
2πi ∫c −i∞ s(s + 1)...(s + m )
φ (y) = ds = 0 if 0 < y ≤ 1 , (65)

where y ∈ R + , m ∈ Z + and c ≥ 1 .
The above equality for the discontinuous factor φ ( y ) is easily verified with the Cauchy residue
theorem.
Hence, there are two cases.

Case 1. 1 ≤ y .

ys
The term (66) is meromorphic with residues
s (s + 1)...(s + m )

y −k
= y−k
(− 1) k
(67)
(− k )(− k + 1)...(− k + k − 1)(− k + k + 1)...(− k + m ) k!(m − k )!

where 0 ≤ k ≤ m . Therefore the sum of these residues is

k m
m
(− 1)k = 1 m  m  − 1  1m − k = 1 1 − 1 
∑y
k =0
−k
∑
k!(m − k )! m! k = 0  k  y  m!  y 
. (68)

Now consider the left contour. The integral along the vertical segment at c in the right-half plane
approaches
c + i∞ ys
∫c −i∞ s(s + 1)...(s + m) ds (69)

27
as T goes to infinity. Along the two horizontal segments from − T ± iT to c ± iT , the integrand is
yσ 1 y1+ σ
bounded by and because the term with σ = −T , σ = c vanishes as T goes to
Tm 1+σ T m
infinity (recall that 1 ≤ y ), the contribution from these two segments is zero. The integrand is
y −T
bounded by on the vertical segment in the left half-plane; hence the integral is
T (T − 1)...(T − m )
2 y −T
bounded by and its contribution is zero also.
(T − 1)...(T − m )
Case 2. 0 < y ≤ 1 .

Consider the contour in the right half-plane. Along the horizontal segments we may use the same
bound as in the first case, with σ = c and σ = T ; hence these integrals vanish (0 < y ≤ 1) . The
yT
integrand is bounded by on the vertical segment in the right half-plane; its
T (T + 1)...(T + m )
contribution is zero because 0 < y ≤ 1 .
The principal feature of the “discontinuous factor” is that it can be used to evaluate finite sums.
Suppose we have a finite sum over the indices k from 1 to n − 1 . Evidently φ ( y ) is non-zero if 1 / y
lies in (0,1) and zero otherwise. We need only find a map such that the set {1,...n − 1} maps to a
subrange of (0,1) and {n, n + 1...} to a subrange of [1, ∞ ) . Clearly 1 / y = k / n is such a map. We
obtain
m
1 c + i∞ 1 ns 1  k
2πi ∫c − i∞ k s (s + 1)...(s + m )
s
ds = 1 −  if k < n
m!  n 
s
1 c + i∞ 1 n
∫
2πi c − i∞ k s(s + 1)...(s + m )
s
ds = 0 if n ≤ k (70)

By a formal argument we finally have

m
1 n −1  k  1 c + i∞ λk  ns
∑ λk 1 −  =
m! k =1  n 
 ∑
2πi ∫c − i∞  k s  s (s + 1)...(s + m )
 ds . (71)

This is the Mellin-Perron formula.

The Mellin-transform view adds two additional perspectives. One, that the Mellin-Perron formula is
a specific instance of harmonic sum formulas, and hence, two, that its evaluation corresponds to
Mellin inversion.
m
 k
We wish to evaluate the harmonic sum ∑
1≤ k < n
λk 1 − 
 n
(72) where m, n ∈ Z + . This is
∞
k 
∑ λ g  n  where g ( x ) = (1 − x ) if 0 < x ≤ 1 ; g ( x ) = 0 otherwise. It
m
equivalent to k (73)
1

1
is no difficult to see that Μ[g ( x ); s ] = . (74)
s (s + 1)...(s + m )
∞
k  ∞
Evidently the sum ∑1 λk g  n  is a harmonic sum G ( x ) of the form ∑1 λk g (kx ) with amplitudes
λk , frequencies µ k = k and evaluated at x = 1 / n . Therefore the transform function G ∗ (s ) is
28
 1
Λ (s ) . (75)
s (s + 1)...(s + m )

By Mellin inversion we thus have

1 c + i∞ x− s
2πi ∫c − i∞
G(x ) = Λ ( s ) ds (76)
s (s + 1)...(s + m )

and in particular

m
1  k 1 c + i∞ ns
G  = ∑ λk 1 −  =
2πi ∫c − i∞
Λ ( s ) ds . (77)
 n  1≤ k < n  n  s (s + 1)...(s + m )

This is the Mellin-Perron formula.

The Mellin transform: Definitions, Theorems and Lemmas

Definition 2.2.1

The open strip of complex numbers α , β is the set {s = σ + it α < σ < β }.

Definition 2.2.2

Let f ( x ) be locally Lebesgue integrable over (0,+∞ ) . The Mellin transform of f ( x ) is defined by

+∞
Μ[ f ( x ); s ] = f ∗ (s ) = ∫ f ( x )x s −1dx . (78)
0

The fundamental strip is the largest open strip where the integral converges.

Lemma 2.2.1

The conditions f ( x )x →0 + ∈ Ο(x u ) , f ( x )x → +∞ ∈ Ο(x v ) , (79)

when u > v , guarantee that f ∗ ( x ) exists in the strip − u,−v .

Definition 2.2.3

Let H 0 ( x ) = 1 if x ∈ [0,1] ; H 0 ( x ) = 0 if x > 1 (80) be defined on [0,+∞ ) and let

H m ( x ) = (1 − x ) H 0 ( x ) when m ∈ Z + . (81)
m

Note that H 0 ( x ) has a discontinuity at x = 1 ; we have lim x →1− H 0 ( x ) = 1 and lim x →1+ H 0 ( x ) = 0 .
Note also that lim x →1− H m ( x ) = lim x →1+ H m ( x ) = 0 when m ∈ Z + ; H m ( x ) is continuous at x = 1 .

Lemma 2.2.2

29
The Mellin transform H m∗ ( x ) of H m ( x ) , where m ∈ N , exists in 0,+∞ and is given by

m!
H m∗ ( x ) = . (82)
s (s + 1)...(s + m )

We have H m ( x )x → 0 + ∈ Ο(1) and H m ( x )x → +∞ ∈ Ο(x − b ) for any b > 0 and for m ∈ N , hence H m∗ ( x )
exists in 0,+∞ . Note that

1 s1 1
1
H 0∗ ( x ) = ∫ x s −1dx =
0 s
[ ]
x 0= .
s
(83)

We also have
1 1 1
H m∗ (s ) = ∫ H m ( x )x s −1dx = ∫ H m −1 ( x )x s −1dx − ∫ H m −1 ( x )x s dx =
0 0 0

= H m∗ −1 ( x ) − ∫
1 (1 − x ) m
sx s −1dx = H m∗ −1 ( x ) −
s ∗
H m (x ) . (84)
0 m m

m
This gives H m∗ ( x ) = H m∗ −1 ( x ) (85)
s+m

Now, we will be concerned with the linearity and the rescaling property of the Mellin transform.

Theorem 2.2.1

Let Κ ⊂ Z be a finite set of integers; let µ k , λk ∈ R + . Let the fundamental strip of Μ[ f ( x ); s ] be

α , β . We have
   λ 
Μ ∑ λk f (µk x ); s  =  ∑ ks Μ[ f ( x ); s ] , (86)
 k   k µk 

where s ∈ α , β .
Let y = µ k x and dy = µ k dx . Note that

∞   ∞ ∞ dy  λ 
∫  ∑ λ f (µ x ) x dx = ∑ λk ∫ f (µk x )x s −1dx = ∑ λk ∫ f ( y ) y s −1 =  ∑ ks  f ∗ (s ) . (87)
s −1
k k s
0
k k
0
k
0 µ  k µk 
k

We were able to exchange the integral with the summation because Κ is finite. It can be shown that
this operation extends to infinite Κ as long as ∑ k λk / µ ks converges absolutely. The extended
property holds in the intersection of the half-plane of convergence of ∑ k
λk / µ ks and the
fundamental strip α , β of f ( x ) .

Definition 2.2.4

1. (Lebesgue integration)

30
Let f ( x ) be integrable with fundamental strip α , β . If c ∈ (α , β ) and f ∗ (c + it ) is integrable, then

1 c + i∞ ∗
2πi ∫c − i∞
f (s )x − s ds = f ( x ) (88)

almost everywhere. If f ( x ) is continuous, the equality holds everywhere on (0,+∞ ) .

2. (Riemann integration.)

Let f ( x ) be locally integrable with fundamental strip α , β and be of bounded variation in a

neighbourhood of x0 . Then

1 c + iT ∗ f x0+ + f x0− ( ) ( )
T → ∞ 2πi ∫c − iT
lim f ( s )x −s
ds x0 = (89)
2

for c ∈ (α , β ) . Of course if lim x → x + f ( x ) = lim x → x − f ( x ) then

0 0

( ) ( )
f x0+ + f x0−
= f ( x0 ) . (90)
2

Theorem 2.2.2 (Mellin-Perron formula)

Let c ∈ R + lie in the half-plane of absolute convergence of ∑ k

λk / k s . Then we have

m
1  k 1 c + i∞ λk  s ds
∑ λk 1 −  =
m! 1≤ k < n  n  ∫  ∑ s n
2πi c − i∞  k ≥1 k  s (s + 1)...(s + m )
(91)

for m ∈ Z + . We have
λn 1 c + i∞ λk  s ds
∑λ
1≤ k < n
k +
2
=  ∑ n
2πi ∫c − i∞  k ≥1 k s  s
(92)

when m = 0 .
This theorem is a straightforward application of Mellin inversion.

Proof.

Let F ( x ) = ∑ k λk f (µ k x ) and use the rescaling property to obtain

 λ 
Μ[F ( x ); s ] = F ∗ (s ) =  ∑ ks  f ∗ (s ) . (93)
 k µk 

Consider Riemann-integrable f ( x ) and apply the Mellin inversion formula

31
 ( ) (
f µk x + + f µk x − )
1 c + iT λk  ∗
∑λ k = lim
2πi ∫  ∑ s  f (s )x − s ds . (94)
k 2 T → ∞ c − iT
 k µk 

Let f ( x ) = H m ( x ) , m ∈ N and let µ k = k . Recall that the fundamental strip of H m ( x ) is 0, ∞ ; let

x = 1 / n . This gives
 k   k 
Hm −  + Hm + 
f (µ x ) + f (µ x )
+ −

∑k λk k 2 k = ∑k λk   2   =
n n

m m
 k   k 
1 − −  + 1 − + 
= ∑ λk  n   n 
+ λn
H m 1+ + H m 1− ( )
 k
m

= ∑ λk 1 −  + λn
( )
H m 1+ + H m 1−
. (95)
( ) ( )
1≤ k < n 2 2 1≤ k < n  n 2

Note that

λn
( )
H m 1+ + H m 1− ( )
= λn / 2 if m = 0 ; λn
( ) ( )
H m 1+ + H m 1−
= 0 if m ∈ Z + . (96)
2 2

Continuing the substitution, we have

1 c + iT λk  ∗ 1 c + iT λk  m!
lim ∫  ∑ s  f (s )x − s ds = lim ∫ ∑ s  n s ds =
T → ∞ 2πi c − iT
 k µk  T → ∞ 2πi c − iT ( ) (
 k k  s s + 1 ... s + m )
m! c + i∞ λk  s ds
= ∫  ∑ s n . (97)
2πi c − i∞
 k k  s (s + 1)...(s + m )

This concludes the proof. Because the fundamental strip of H m ( x ) is 0, ∞ , the choice of c > 0 is
determined by the half-plane of convergence of ∑ k
λk / k s only.
Now we presents two Mellin-Perron formulae for the generalized ζ -function. We apply the Mellin
inversion theorem to F ( x ) = ∑ k λk f (µ k x ) with x = r / n , r, n ∈ Z + , µ k = k + a , λk = 1 , a ∈ R ,
a ∈ (0,1] , f ( x ) = H1 ( x ) = (1 − x )H 0 (x ) . As we require µ k ∈ R + we take k ∈ N . We have

 r  r
F (x ) = ∑ 1 − (k + a ) n  H  (k + a ) n 
k∈N
0 (98)

 1  ∗ ζ (s , a )
and F ∗ (s ) =  ∑  f (s ) =
s 
(99)
 k ∈N (k + a )  s (s + 1)

where σ > 1 . We need to evaluate F ( x ) . H 0 ( x ) vanishes outside of [0,1) , hence we require

0 ≤ (k + a )r / n < 1 or k < n / r − a . Let N (u ) = {v < u v ∈ N } where u ∈ R + . We have

 r
F (x ) = ∑
(
k∈N
1 − (k + a )  . (100)
n / r −a ) n

With these settings the Mellin inversion formula yields the following theorem.

32
Theorem 2.2.3

Let c > 1 .
 r  1 c + i∞ 1 ns
∑ 
k∈N (n / r − a ) 
1 − (k + a )
n  2πi ∫c − i∞ r s
 = ζ ( s , a )
s (s + 1)
ds . (101)

This theorem has several useful corollaries. The first of these is obtained by setting r = 1 . Let
α ∈ (− 1,0) .

Corollary 2.2.3

Let n ∈ N .
1 α + i∞ ns
2πi ∫α − i∞
ζ ( s , a ) ds = 0 . (102)
s (s + 1)

Let c = 1 . The set of poles of ζ (s, a )n s / (s (s + 1)) in α , c is {1,0} . We apply the shifting lemma
with Φ(s ) = n s and T j = j . Because n s = nσ we can take M = n c .

1 α + i∞ ns 1 c + i∞ ns
2πi ∫α − i∞ 2πi ∫c − i∞
ζ (s , a ) ds = ζ (s , a )
s (s + 1) s(s + 1)
 ns   ns 
− Re s ζ (s, a )
 ; s = 1 − Re s ζ (s, a )
  ; s = 0 
 s (s + 1)   s (s + 1) 
 1 n a 11
= ∑ 1 − (k + a )  − − ζ (0, a ) = n − n − (n − 1)n − n − ζ (0, a ) = 1 − a − ζ (0, a ) = 0 . (103)
0≤k <n  n 2 n n2 2 2

The second corollary results from taking r = 4 .

Corollary 2.2.4

Let n ∈ N .
1 α + i∞ 1 ns
2πi ∫α − i∞ 4 s
ζ ( s , a ) ds . (104)
s(s + 1)

( )
We let c = 1 as before and consider the poles of ζ (s, a )n s / 4 s s (s + 1) in α , c , which are at 1 and
0. We apply the shifting lemma with Φ(s ) = (n / 4 ) , T j = j and take M = (n / 4 ) .
s c

1 α + i∞ ns 1 c + i∞ ns
2πi ∫α − i∞ 2πi ∫c − i∞
ζ ( s , a ) ds = ζ ( s , a )
4 s s (s + 1) 4 s s (s + 1)
 ns   ns 
− Re s ζ (s, a ) s ; s = 1 − Re s ζ (s, a ) s ; s = 0 
 4 s (s + 1)   4 s (s + 1) 
 4 n n
= ∑ 1 − (k + a )  − − ζ (0, a ) = ε (n, a ) − − ζ (0, a ) . (105)
k∈N (n / 4 − a )  n 8 8

33
Suppose n = 4m + m1 where m1 ∈ {0,1,2,3}. We have n / 4 − a = [n / 4] + m1 / 4 − a . If m1 / 4 < a , the
sum over N (n / 4 − a ) ranges from 0 to [n / 4] − 1 . If m1 / 4 ≥ a the sum includes [n / 4] . We have two
cases:
n 4  n  2   n   n  m
ε (n, a ) =   − a   −    − 1   if 1 < a
4 n  4  n   4   4  4
n 4 n  2 n  n  m1
ε (n, a ) =   + 1 − a    + 1 −    + 1   if ≥a.
4 n 4  n 4
 4
  4

We note that [n / 4] = (n − m1 ) / 4 and [n / 4]4 / n = 1 − m1 / n . Hence the two terms evaluate to

1 1 1 1 1  1 1 1 1 1 
n + − a +  am1 − m1 − m12  and n + − a +  a (m1 − 4 ) + m1 − m12  .
8 2 n 2 8  8 2 n 2 8 

We conclude that

1 α + i∞ ns 1 1
2πi ∫α − i ∞
ζ ( s , a )
4 s (s + 1)
s
ds = ε (n, a ) − n − + a .
8 2
(106)

2.3 The zeta-function quantum field theory and the quantum L-functions.[6]

The Riemann zeta-function is defined as

∞
1
ζ (s ) = ∑ s
, s = σ + iτ , σ > 1 (107)
n =1 n

and there is an Euler adelic representation

ζ (s ) = ∏ (1 − p −s ) . (108)
−1

Now, we have the Riemann ξ -function

s (s − 1) − 2  s 
s
ξ (s ) = π Γ ζ (s ) (109)
2 2

which is an entire function. The zeros of the ξ -function are the same as the nontrivial zeros of the
ζ -function. There is the functional equation

ξ (s ) = ξ (1 − s ) (110)

and the Hadamard representation for the ξ -function

1  s  s/ρ
ξ (s ) = e as ∏ 1 − e . (111)
2 ρ  ρ 

34
Here ρ are nontrivial zeros of the zeta-function and

1 1
a = − γ − 1 + log 4π (112)
2 2

where γ is Euler’s constant.

If F (τ ) is a function of a real variable τ then we define a pseudo-differential operator F ( ) by
using the Fourier transform
~
( )
F ( ) φ ( x ) = ∫ eixk F k 2 φ (k )dk . (113)

Here is the d’Alambertian operator

∂2 ∂2 ∂2
=− + + ... + , (114)
∂x02 ∂x12 ∂xd2−1

~
φ (x ) is a function from x ∈ R d , φ (k ) is the Fourier transform and k 2 = k02 − k12 − ... − k d2−1 .
We assume that the integral (113) converges.
1 
One can introduce a natural field theory related with the real valued function F (τ ) = ξ  + iτ 
 2 
defined by means of the zeta-function. We consider the following Lagrangian

L = φξ (1 / 2 + i ) φ , (115)

the integral
~
ξ (1 / 2 + i ) φ ( x ) = ∫ eixkξ (1 / 2 + i ) φ (k )dk (116)

1 
converges if φ ( x ) is a decreasing function since ξ  + iτ  is bounded.
2 
The operator ξ (1 / 2 + i ) (or ζ (1 / 2 + i )) is the first quantization the Riemann zeta-function.
From the Hadamard representation (111) we get

1  C ∞  τ2 
ξ  + iτ  =
2 
∏ 1 −  . (117)
2 n=1  mn4 

It is possible to write the formula (117) in the form

1  C  τ 
ξ  + iτ  = ∏ 1 + 2  (118)
2  2 ε ,n  εmn 

where ε = ±1 and a regularization is assumed.

To quantize the zeta-function classical field φ ( x ) which satisfies the equation in the Minkowski
space
F ( ) φ ( x ) = 0 (119)

35
where F ( ) = ξ (1 / 2 + i ) we can try to interpret φ ( x ) as an operator valued distribution in a
Hilbert space Η which satisfies the equation (119). We suppose that there is a representation of the
Poincare group and an invariant vacuum vector 0 in Η . Then the Wightman function

W ( x − y ) = 0 φ ( x )φ ( y ) 0

is a solution of the equation

F ( ) W (x ) = 0 . (120)

By using (118) we can write the formal Kallen-Lehmann representation

(
W ( x ) = ∑ ∫ eixk f εn (k )δ k 2 + εmn2 dk . ) (121)
εn

One introduces also another useful function

 1 iτ 
Γ + 
Z (τ ) = π −iτ / 2 4 2  ζ  1 + iτ  = eiϑ (τ )ζ  1 + iτ  .
    (122)
 1 iτ  2  2 
Γ + 
4 2 

Here Γ( z ) is the gamma function. The function Z (τ ) is called the Riemann-Siegel (or Hardy)
function. It is known that Z (τ ) is real for real τ and there is a bound

Z (τ ) = O τ ( ), ε
ε > 0 . (123)

One can introduce a natural field theory related with the real valued functions Z (τ ) defined by
means of the zeta-function by considering the following Lagrangian

L = φZ ( ) φ .

The integral (113) converges if φ ( x ) is a decreasing function since there is the bound (123).
Thence, we have the following connection:

 1 iτ 
Γ + 
π −iτ / 2 4 2
 1 iτ
 ζ  1 + iτ  = e iϑ (τ )ζ  1 + iτ  = O τ ε ⇒
 2
 


2


( )
Γ + 
4 2 

~
( )
⇒ F ( ) φ ( x ) = ∫ eixk F k 2 φ (k )dk , ε > 0 . (124)

For any character to modulus q one defines the corresponding Dirichlet L-function by setting

∞
χ (n )
L (s , χ ) = ∑ , (σ > 1) . (125)
n =1 ns

36
If χ is primitive then L(s, χ ) has an analytic continuation to the whole complex plane. The zeros
lie in the critical strip and symmetrically distributed about the critical line σ = 1 / 2 .
If we quantize the L-function by considering the pseudo-differential operator

L(σ + i , χ ) (126)

then we can try to avoid the appearance of tachyons and/or ghosts by choosing an appropriate
character χ .
The Taniyama-Weil conjecture relates elliptic curves and modular forms. It asserts that if E is an
elliptic curve over Q , then there exists a weight-two cusp form f which can be expressed as the
Fourier series
f ( z ) = ∑ an e 2πnz (127)

with the coefficients an depending on the curve E. Such a series is a modular form if and only if its
Mellin transformation, i.e. the Dirichlet L-series

L(s, f ) = ∑ an n − s (128)

has a holomorphic extension to the full s-plane and satisfies a functional equation. For the elliptic
curve E we obtain the L-series L(s, E ) . The Taniyama-Weil conjecture was proved by Wiles and
Taylor for semistable elliptic curves and it implies Fermat’s Last Theorem.
Quantization of the L-functions can be performed similarly to the quantization of the Riemann zeta-
function discussed above by considering the corresponding pseudo-differential operator L(σ + i ).

Chapter 3.

How primes and adeles are related to the Riemann zeta function[7]

A. Connes has reduced the Riemann hypothesis for L-function on a global field k to the validity of a
trace formula for the action of the idele class group on the noncommutative space quotient of the
adeles of k by the multiplicative group of k.
Connes has devised a Hermitian operator whose eigenvalues are the Riemann zeros on the critical
line. Connes gets a discrete spectrum by making the operator act on an abstract space where the
primes appearing in the Euler product for the Riemann zeta function are built in; the space is
constructed from collections of p-adic numbers (adeles) and the associated units (ideles).

37
Hence, the geometric framework involves the space X of Adele classes, where two adeles which
belong to the same orbit of the action of GL1 (k ) ( k a global field), are considered equivalent. The
group Ck = GL1 ( A) / GL1 (k ) of Idele classes (which is the class field theory counterpart of the Galois
group) acts by multiplication on X.
We have a trace formula (Theorem 3) for the action of the multiplicative group K ∗ of a local field
K on the Hilbert space L2 (K ) , and (Theorem 4) a trace formula for the action of the multiplicative
group Cs of Idele classes associated to a finite set S of places of a global field k , on the Hilbert
space of square integrable functions L2 ( X S ) , where X S is the quotient of ∏ v∈S
kv by the action of
the group OS∗ of S-units of k . The validity of the trace formula for any finite set of places follows
from Theorem 4, but in the global case is left open and shown (Theorem 5) to be equivalent to the
validity of the Riemann Hypothesis for all L functions with Grossencharakter.

H. Montgomery has proved (assuming RH) a weakening of the following conjecture (with
α , β > 0 ),
  sin (πu )  2 
Card {(i, j ); i, j ∈ 1,..., M ; xi − x j ∈ [α , β ]} ≈ M ∫ 1 − 
β

α
 du (1)
  π u  

This law, i.e. the equation (1), is precisely the same as the correlation between eigenvalues of
hermitian matrices of the gaussian unitary ensemble. Moreover, numerical tests due to A. Odlyzko
have confirmed with great precision the behaviour (1) as well as the analogous behaviour for more
than two zeros. N. Katz and P. Sarnak has proved an analogue of the Montgomery-Odlyzko law for
zeta and L-functions of function fields over curves.
It is thus an excellent motivation to try and find a natural pair (Η, D ) where naturality should mean
for instance that one should not even have to define the zeta function, let alone its analytic
continuation, in order to obtain the pair (in order for instance to avoid the joke of defining Η as the
l 2 space built on the zeros of zeta).

Theorem 1.

( )
Let K be a local field with basic character α . Let h ∈ S K ∗ have compact support. Then RΛU (h )
is a trace class operator and when Λ → ∞ , one has

h(u '
−1
) d u + o(1)
Trace(RΛU (h )) = 2h(1)log' Λ + ∫ ∗
(2)
1− u

'
where 2 log' Λ = ∫ d ∗λ , and the principal value ∫ is uniquely determined by the pairing
∗
[ −1
λ ∈K , λ ∈ Λ , Λ ]
du
with the unique distribution on K which agrees with for u ≠ 1 and whose Fourier transform
1− u
vanishes at 1.

Proof.

We normalize the additive Haar measure to be the selfdual one on K . Let the constant ρ > 0 be
determined by the equality,

38
 dλ
∫1≤ λ ≤ Λ λ
≈ ρ log Λ when Λ → ∞ , (3)

dλ du
so that d ∗λ = ρ −1 . Let L be the unique distribution, extension of ρ −1 whose Fourier
λ 1− u
transform vanishes at 1, Lˆ (1) = 0 . One then has by definition,

' ( )
h u −1 ∗ h u −1 ( )
∫ 1− u d u = L ,
u
, (4)

where
( )
h u −1
= 0 for u −1 outside the support of h . Let T = U (h ) . We can write the Schwartz
u
kernel of T as,
( )
k ( x, y ) = ∫ h λ−1 δ ( y − λx )d ∗λ . (5)

Given any such kernel k we introduce its symbol,

σ ( x, ξ ) = ∫ k ( x, x + u )α (uξ )du (6)

as its partial Fourier transform. The Schwartz kernel rΛt ( x, y ) of the transpose RΛt is given by,

rΛt ( x, y ) = ρ Λ (x )(ρˆ Λ )( x − y ) . (7)

Thus, the symbol σ Λ of RΛt is simply,

σ Λ ( x, ξ ) = ρ Λ ( x )ρ Λ (ξ ) . (8)

The operator RΛ is of trace class and one has,

Trace (RΛT ) = ∫ k ( x, y )rΛt ( x, y )dxdy . (9)

Using the Parseval formula we thus get,

Trace (RΛT ) = ∫ σ ( x, ξ )dxdξ . (10)

x ≤ Λ, ξ ≤Λ

Now the symbol σ of T is given by,

( )
σ ( x, ξ ) = ∫ h(λ−1 ) ∫ δ ( x + u − λx )α (uξ )du d ∗λ . (11)

One has,
∫ δ (x + u − λx )α (uξ )du = α ((λ − 1)xξ ) , (12)

thus (11) gives,

39
 σ ( x, ξ ) = ρ −1 ∫ g (λ )α (λxξ )dλ (13)
K

where,
(
g (λ ) = h (λ + 1) λ + 1 .
−1
) −1
(14)

Since h is smooth with compact support on K ∗ the function g belongs to Cc∞ (K ) . Thus
σ ( x, ξ ) = ρ −1 gˆ ( xξ ) and
Trace (RΛT ) = ρ −1 ∫ gˆ ( xξ )dxdξ . (15)
x ≤ Λ, ξ ≤Λ

dx
With u = xξ one has dxdξ = du and, for u ≤ Λ2 ,
x

dx
ρ −1 ∫ u = 2 log' Λ − log u . (16)
Λ
≤ x ≤Λ x

Thus we can rewrite (15) as,

Trace (RΛT ) = ∫ gˆ (u )(2 log' Λ − log u )du . (17)

u ≤ Λ2

Since g ∈ Cc∞ (K ) one has,

∫ u ≥ Λ2
gˆ (u ) du = O Λ− N ( ) ∀N (18)

and similarly for gˆ (u )log u . Thus

Trace (RΛT ) = 2 g (0 )log' Λ − ∫ gˆ (u )log u du + o(1) . (19)

Now for any local field K and basic character α , if we take for the Haar measure da the selfdual
one, the Fourier transform of the distribution ϕ (u ) = − log u is given outside 0 by

1
ϕˆ (a ) = ρ −1 , (20)
a

with ρ determined by (3). To see this one lets P be the distribution on K given by,

P ( f ) = lim  ∫ f ( x )d ∗ x + f (0 )log ε  . (21)

ε →0
ε ∈Mod ( K )
 x ≥ε 

One has P( f a ) = P( f ) − log a f (0) which is enough to show that the function P̂ ( x ) is equal to
− log x + cst, and ϕ̂ differs from P by a multiple of δ 0 . Thus the Parseval formula gives, with the
convention of Theorem 3,

40
 1 ' da
− ∫ gˆ (u )log u du =
ρ∫
g (a ) . (22)
a

Replacing a by λ − 1 and applying (14) gives the desired result.

Now, let k be a global field and S a finite set of places of k containing all infinite places. The
{
group OS∗ of S -units is defined as the subgroup of k ∗ , OS∗ = q ∈ k ∗ , qv = 1, v ∉ S . It is co-compact }
in J S1 where, J S = ∏ kv∗ and, J S1 = { j ∈ J S , j = 1}. Thus the quotient group CS = J S / OS∗ plays
v∈S

the same role as Ck , and acts on the quotient X S of AS = ∏v∈S kv by OS∗ .

Theorem 2.

Let AS be as above, with basic character α = ∏ α v . Let h ∈ S (CS ) have compact support. Then
when Λ → ∞ , one has

Trace (RΛU (h )) = 2h(1)log' Λ + ∑ ∫ ∗

h u −1 ∗
d u + o(1)
' ( ) (23)
v∈ S v 1 − u
k

where 2 log' Λ = ∫ ∗
, each kv∗ is embedded in CS by the map u → (1,1,..., u,...,1) and
[
λ ∈C S , λ ∈ Λ−1 , Λ ]d λ
'
the principal value ∫ is uniquely determined by the pairing with the unique distribution on kv
du
which agrees with for u ≠ 1 and whose Fourier transform relative to α v vanishes at 1.
1− u

Proof.

We normalize the additive Haar measure dx to be the selfdual one on the abelian group AS . Let the
constant ρ > 0 be determined by the equality,

dλ
∫λ ∈D ,1≤ λ ≤ Λ λ
≈ ρ log Λ when Λ → ∞ ,

dλ
so that d ∗λ = ρ −1 . We let f be a smooth compactly supported function on J S such that
λ

∑ f (qg ) = h(g ) ∀g ∈ CS . (24)

q∈O S∗

The existence of such an f follows from the discreteness of OS∗ in J S . We then have the equality
U ( f ) = U (h ) , where
U ( f ) = ∫ f (λ )U (λ )d ∗λ . (25)

41
Now, for an operator T, acting on functions on AS , which commutes with the action of OS∗ and is
represented by an integral kernel,

T (ξ ) = ∫ k ( x, y )ξ ( y )dy , (26)

the trace of its action on L2 ( X S ) is given by,

Tr (T ) = ∑ ∫ k (x, qx )dx ,
D
(27)
q∈O S∗

where D is a fundamental domain for the action of OS∗ on the subset J S of AS , whose complement
is negligible. Let T = U ( f ) . We can write the Schwartz kernel of T as,

( )
k ( x, y ) = ∫ f λ−1 δ ( y − λx )d ∗λ , (28)

by construction one has,

k (qx, qy ) = k ( x, y ) q ∈ OS∗ . (29)

For any q ∈ OS∗ , we shall evaluate the integral,

Iq = ∫ k (qx, y )rΛt ( x, y )dydx (30)

x∈D

where the Schwartz kernel rΛt ( x, y ) for the transpose RΛt is given by,

rΛt ( x, y ) = ρ Λ (x )(ρˆ Λ )( x − y ) . (31)

To evaluate the above integral, we let y = x + a and perform a Fourier transform in a . For the
Fourier transform in a of rΛt ( x, x + a ) , one gets,

σ Λ ( x, ξ ) = ρ Λ ( x )ρ Λ (ξ ) . (32)

For the Fourier transform in a of k (qx, x + a ) , one gets,

(
σ ( x, ξ ) = ∫ f (λ−1 ) ∫ δ ( x + a − λqx )α (aξ )da d ∗λ . (33) )
One has,
∫ δ (x + a − λqx )α (aξ )da = α ((λq − 1)xξ ) , (34)

thus (33) gives,

σ ( x, ξ ) = ρ −1 ∫ g q (u )α (uxξ )du (35)
AS

where,
(
g q (u ) = f q (u + 1) u + 1 .
−1
) −1
(36)

42
Since f is smooth with compact support on AS∗ the function g q belongs to Cc∞ ( AS ) .
Thus σ ( x, ξ ) = ρ −1 gˆ q (xξ ) and, using the Parseval formula we get,

Iq = ∫ σ (x, ξ )dxdξ . (37)

x∈D , x ≤ Λ , ξ ≤ Λ

This gives,
I q = ρ −1 ∫ gˆ q ( xξ )dxdξ . (38)
x∈D , x ≤ Λ , ξ ≤ Λ

dx
With u = xξ one has dxdξ = du and, for u ≤ Λ2 ,
x

dx
ρ −1 ∫ u = 2 log' Λ − log u . (39)
x∈ D , ≤ x ≤ Λ x
Λ

Thus we can rewrite (38) as,

Trace (RΛT ) = ∑∫ gˆ q (u )(2 log' Λ − log u )du . (40)

u ≤ Λ2
q∈O S∗

Now log u = ∑v∈S log uv , and we shall first prove that,

∑ ∫ gˆ (u )du = h(1) ,
q (41)
q∈O S∗

while for any v ∈ S ,

( )
h u −1 ∗
∑ ∫ gˆ q (u )(− log uv )du = ∫ ∗
'
d u . (42)
q∈O S∗
kv 1− u

In fact all the sums in q will have only finitely many non zero terms. It will then remain to control
the error term, namely to show that,

∑ ∫ gˆ (u )(log u − 2 log' Λ ) du = 0(Λ ),

+ −N
q (43)
q∈O S∗

for any N , where we used the notation x + = 0 if x ≤ 0 and x + = x if x > 0 .

(−1
)
Now recall that for (36), g q (u ) = f q (u + 1) u + 1 , so that ∫ gˆ q (u )du = g q (0 ) = f (q ) . Since f
−1

has compact support in AS∗ , the intersection of OS∗ with the support of f is finite and by (24) we
get the equality (41). To prove (42), we consider the natural projection prv from ∏ l ∈S
kl∗ to

∏ l ≠v
kl∗ . The image prv (OS∗ ) is still a discrete subgroup of ∏ l ≠v
kl∗ , thus there are only finitely
many q ∈ OS∗ such that kv∗ meets the support of f q , where f q (a ) = f (qa ) for all a .

43
For each q ∈ OS∗ one has,
( )d u ,
f q u −1
∫ gˆ q (u )(− log uv )du = ∫ ∗
'
∗
(44)
kv 1− u

and this vanishes except for finitely many q ' s , so that by (24) we get the equality (42).

Theorem 3.

Let k be a global field of positive characteristic and QΛ be the orthogonal projection on the
subspace of L2 ( X ) spanned by the f ∈ S ( A) such that f ( x ) and fˆ ( x ) vanish for x > Λ . Let
h ∈ S (Ck ) have compact support. Then the following conditions are equivalent,

a) When Λ → ∞ , one has

' h(u −1 ) ∗
Trace (QΛU (h )) = 2h(1)log' Λ + ∑ ∫ ∗ d u + o(1) . (45)
v
kv 1− u

b) All L functions with Grossencharakter on k satisfy the Riemann Hypothesis.

To prove that (a) implies (b), we shall prove (assuming (a)) the positivity of the Weil distribution,

∆ = log d −1 δ1 + D − ∑ Dv . (46)
v

We have that for δ = 0 , the map E ,

E ( f )(g ) = g ∑ f (qg )
1/ 2
∀g ∈ Ck , (47)
q∈k ∗

defines a surjective isometry from L2 ( X )0 to L2 (Ck ) such that,

EU (a ) = a V (a )E ,
1/ 2
(48)

where the left regular representation V of Ck on L2 (Ck ) is given by,

(V (a )ξ )(g ) = ξ (a −1g ) ∀g , a ∈ Ck . (49)

Let S Λ be the subspace of L2 (Ck ) given by,

{
S Λ = ξ ∈ L2 (Ck ); ξ ( g ) = 0, ∀g , g ∉ Λ−1 , Λ . (50) [ ]}
We shall denote by the same letter the corresponding orthogonal projection.

44
Let BΛ ,0 be the subspace of L2 ( X )0 spanned by the f ∈ S ( A)0 such that f ( x ) and fˆ ( x ) vanish for
x > Λ and QΛ ,0 be the corresponding orthogonal projection. Let f ∈ S ( A)0 be such that f ( x ) and
fˆ ( x ) vanish for x > Λ , then E ( f )( g ) vanishes for g > Λ , and the equality

1
()
E ( f )( g ) = E fˆ   f ∈ S ( A)0 , (51)
g

shows that E ( f )( g ) vanishes for g < Λ−1 .

This shows that E (BΛ ,0 ) ⊂ S Λ , so that if we let QΛ' , 0 = EQΛ , 0 E −1 , we get the inequality,

QΛ' , 0 ≤ S Λ (52)

and for any Λ the following distribution on Ck is of positive type,

(( )
∆ Λ ( f ) = Trace S Λ − QΛ' , 0 V ( f ) , (53))
i.e. one has,
( )
∆Λ f ∗ f ∗ ≥ 0 , (54)

( )
where f ∗ ( g ) = f g −1 for all g ∈ Ck .
Let then f ( g ) = g
−1 / 2
( ) ~ ~
() ( )
h g −1 , so that by (48) one has EU (h ) = V f E where f ( g ) = f g −1 for all
g ∈ Ck . Then, we have:

( )
h u −1 ∗
∑ D ( f ) − log d = ∑∫∗
'
−1
v d u. (55)
v v
kv 1− u

One has Trace (S ΛV ( f )) = 2 f (1) log' Λ , thus using (a) we see that the limit of ∆ Λ when Λ → ∞ is
the Weil distribution ∆ . The term D in the latter comes from the nuance between the subspaces BΛ
and BΛ ,0 . This shows using (53), that the distribution ∆ is of positive type so that (b) holds.
Let us now show that (b) implies (a). We shall compute from the zeros of L -functions and
independently of any hypothesis the limit of the distributions ∆ Λ when Λ → ∞ .
We choose an isomorphism
Ck ≅ Ck ,1 × N . (56)

where N = range ⊂ R+∗ , N ≅ Z is the subgroup q Z ⊂ R+∗ . For ρ ∈ C we let dµ ρ ( z ) be the

harmonic measure of ρ with respect to the line iR ⊂ C . It is a probability measure on the line iR
and coincides with the Dirac mass at ρ when ρ is on the line.
The implication (b) ⇒ (a) follows immediately from the explicit formulas and the following lemma,

Lemma 1.

The limit of the distributions ∆ Λ when Λ → ∞ is given by,

45
  1 
∆∞ ( f ) =
 1
∑ 
N  χ~, + ρ  ∫ fˆ (χ~, z )dµ ρ ( z )
 2  z∈iR
(57)
L  χ~ , + ρ =0
 2 
ρ∈B / N ⊥

  − 1 1   1 
where B is the open strip B =  ρ ∈ C ; Re(ρ ) ∈  , , N  χ~, + ρ  is the multiplicity of the
  2 2   2 
zero, dµ ρ ( z ) is the harmonic measure of ρ with respect to the line iR ⊂ C , and the Fourier
transform fˆ of f is defined by

fˆ (χ~, ρ ) = ∫ f (u )χ~(u ) u d ∗u . (58)

ρ
Ck

Let us first recall the Weil explicit formulas. One lets k be a global field. One identifies the
quotient Ck / Ck ,1 with the range of the module,

N = {g ; g ∈ Ck } ⊂ R+∗ . (59)

One endows N with its normalized Haar measure d ∗ x . Given a function F on N such that, for
1
some b > ,
2

( )
F (ν ) = 0 ν b ν →0, ( )
F (ν ) = 0 ν − b , ν → ∞ , (60)

one lets,
Φ(s ) = ∫ F (ν )ν 1/ 2− s d ∗ν . (61)
N

Given a Grossencharakter χ , i.e. a character of Ck and any ρ in the strip 0 < Re(ρ ) < 1 , one lets
N (χ , ρ ) be the order of L(χ , s ) at s = ρ . One lets,

S (χ , F ) = ∑ N (χ , ρ )Φ (ρ ) (62)
ρ

where the sum takes place over ρ ' s in the above open strip. One then defines a distribution ∆ on
Ck by,
∆ = log d −1 δ1 + D − ∑ Dv , (63)
v

−1
where δ1 is the Dirac mass at 1 ∈ Ck , where d is a differential idele of k so that d is up to sign
the discriminant of k when char (k ) = 0 and is q 2 g −2 when k is a function field over a curve of
genus g with coefficients in the finite field Fq . The distribution D is given by,

D ( f ) = ∫ f (w ) w
Ck
( 1/ 2
+w
−1 / 2
)d w
∗
(64)

46
where the Haar measure d ∗ w is normalized. The distributions Dv are parametrized by the places v
of k and are obtained as follows. For each v one considers the natural proper homomorphism,

kv∗ → Ck , x → class of (1,..., x,1...) (65)

of the multiplicative group of the local field kv in the idele class group Ck . One then has,

f (u ) 1/ 2 ∗
Dv ( f ) = Pfw∫ ∗ u d u (66)
kv 1− u

where the Haar measure d ∗u is normalized, and where the Weil Principal value Pfw of the integral
is obtained as follows, for a local field K = kv ,

1
Pfw∫ ∗ 1R∗ d ∗u = 0 , (67)
kv v
1− u

if the local field kv is non Archimedean, and otherwise:

Pfw∫ ∗ ϕ (u )d ∗u = PF0 ∫ ∗ ψ (ν )d ∗ν , (68)

kv R+

where ψ (ν ) = ∫ ϕ (u )dν u is obtained by integrating ϕ over the fibers, while

u =ν

(( )
PF0 ∫ψ (ν )d ∗ν = 2 log(2π )c + lim ∫ 1 − f 02t ψ (ν )d ∗ν − 2c log t ,
t →∞
) (69)

where one assumes that ψ − cf1−1 is integrable on R+∗ , and (

f 0 (ν ) = inf ν 1/ 2 ,ν −1/ 2 ) ∀ν ∈ R+∗ ,
f1 = f 0−1 − f 0 . The Weil explicit formula is then,

Theorem 4.

With the above notations one has S (χ , F ) = ∆(F ( w )χ (w)) .

Let K be non Archimedean, furthermore, let α be a character of K such that,

α / R = 1 , α / π −1 R ≠ 1 . (70)

Then, for the Fourier transform given by,

(Ff )(x ) = ∫ f ( y )α ( y )dy , (71)

with dy the selfdual Haar measure, one has

F (1R ) = 1R . (72)

Lemma 2.

47
With the above choice of α one has

' ( )
h u −1 ∗ h u −1 ∗ ( )
∫ 1− u d u = Pfw ∫ 1− u d u (73)

with the notations of theorem 1.

By construction the two sides can only differ by a multiple of h(1) . Let us recall from theorem 1
that the left hand side is given by

L,
( )
h u −1
, (74)
u

du
where L is the unique extension of ρ −1 whose Fourier transform vanishes at 1, Lˆ (1) = 0 .
1− u
Thus from (67) we just need to check that (74) vanishes for h = 1R∗ , i.e. that

L,1R∗ = 0 . (75)

Equivalently, if we let Y = {y ∈ K ; y − 1 = 1} we just need to show, using Parseval, that,

log u ,1̂Y = 0 . (76)

One has 1̂Y ( x ) = ∫ α ( xy )dy = α ( x )1̂R∗ ( x ) , and 1R∗ = 1R − 1P , 1̂R∗ = 1R − π 1π −1R , thus, with q −1 = π ,
Y

 1 
1̂Y ( x ) = α ( x )1R − 1π −1R ( x ) . (77)
 q 

We now need to compute ∫ log x 1̂ (x )dx = A + B ,

Y

1  1
A=− ∫ −1 ∗
q π R
α ( x )(log q )dx , B = 1 −  ∫ log x dx .
 q R
(78)

Let us show that A + B = 0 . One has ∫ dx = 1 ,

R
and

R
( ) ( ( )
A = − ∫ ∗ α π −1 y (log q )dy = − log q ∫ α π −1 y dy − ∫ dy =
R P
) 1
q
log q , (79)

∫ α (π y )dy = 0
−1
since as α / π −1 R ≠ 1 .
R

 1
To compute B , note that ∫π n ∗
R
dy = q −n 1 −  so that
 q

48
 2
 1  1 ∞
1
B = 1 −  ∫ log x dx = 1 −  ∑ (− n log q )q −n
= − q −1 log q = − log q , (80)
 q R
 q n =0 q

and A + B = 0 .
Let us now treat the case of Archimedean fields. We take K = R first, and we normalize the Fourier
transform as,
(Ff )(x ) = ∫ f ( y )e −2πixy dy (81)

so that the Haar measure dx is selfdual.

With the notations of (68) one has,

1/ 2
u
Pfw∫ ∗ f (u )
0
3
d ∗u = log π + γ (82)
R 1− u

where γ is Euler’s constant, γ = −Γ' (1) . Indeed integrating over the fibers gives (
f 04 × 1 − f 04 )
−1
,
and one gets,

( ) d ∗u = log(2π ) + lim ∫ ∗ 1 − f 02t f 04 1 − f 04

d ∗u − log t  = log 2π + γ − log 2 .
( ) ( )
−1 −1
PF0 ∫ ∗ f 04 × 1 − f 04
R+  t →∞  R+ 
(83)
Now let ϕ (u ) = − log u , it is a tempered distribution on R and one has,

2 1 γ
ϕ , e −πu = log π + + log 2 , (84)
2 2

∂  2  1 − s 
s −1
∂ − s −πu 2

∂s ∫
as one obtains from u e du = π Γ  evaluated at s =0, using
∂s   2 
1
Γ'  
 2  = −γ − 2 log 2 . Thus by the Parseval formula one has,
1
Γ 
2

2 1 γ
ϕˆ , e −πx = log π + + log 2 , (85)
2 2

which gives, for any test function f ,

ϕˆ , f = lim  ∫ f ( x )d ∗ x + (log ε ) f (0 ) + λf (0 ) (86)

ε →0 
 x ≥ε 

where λ = log(2π ) + γ . In order to get (86) one uses the equality,

49
  
lim  ∫ f ( x )d ∗ x + (log ε ) f (0 ) = lim  ∫ f ( x ) x d ∗ x − f (0 ) ,
ε 1
(87)
ε →0 
 x ≥ε  ε →0  ε 

which holds since both sides vanish for f ( x ) = 1 if x ≤ 1 , f ( x ) = 0 otherwise. Thus from (86) one
gets,
' 1  f (u ) ∗ 
∫R f (u ) 1 − u d u = λf (1) + lim ∫ d u + (log ε ) f (1) . (88)
∗

ε →0 1−u ≥ε 1 − u
 

f (u ) = u f 03 ( u ) , the right hand side of (88) gives

1/ 2
Taking λ − log 2 = log π + γ , thus we
conclude using (82) that for any test function f ,

' 1 1
∫ f (u ) 1 − u d u = Pfw∫ f (u ) 1 − u d u .
∗ ∗
(89)
R R

Let us finally consider the case K = C . We choose the basic character α as

α ( z ) = exp 2πi ( z + z ) , (90)

the selfdual Haar measure is dzdz = dz ∧ dz , and the function f ( z ) = exp− 2π z is selfdual.
2

The normalized multiplicative Haar measure is

dz ∧ dz
d ∗z = 2
. (91)
2π z

Let us compute the Fourier transform of the distribution

ϕ (z ) = − log z C = −2 log z . (92)

One has
2
ϕ , exp− 2π z = log 2π + γ , (93)

as is seen using
∂
∂ε
(∫ e − 2π z
2

z
−2 ε
dz ∧ dz = ) ∂∂ε [(2π ) Γ(1 − ε )].
ε

2
Thus ϕˆ , exp− 2π u = log 2π + γ and one gets,

ϕˆ , f = lim  ∫ f (u )d ∗u + log εf (0 ) + λ ' f (0 )


(94)
ε →0  u C ≥ε

where λ ' = 2(log 2π + γ ) .

To see this one uses the analogue of (87) for K = C , to compute the right hand side of (94) for
f ( z ) = exp− 2π z . Thus, for any test function f , one has,
2

50
 ' 1  f (u ) ∗ 
∫ f (u )
1− u C
d ∗u = λ ' f (1) + lim  ∫
ε →0 1−u C ≥ε 1 − u
d u + (log ε ) f (1) . (95)
 
C
C

C
−1
Let us compare it with Pfw . When one integrates over the fibers of C ∗ → R+∗ the function 1 − z C
one gets,
1 2π 1 1 1
2π ∫0
1 − eiθ z
2
dθ =
1− z
2
if z < 1 , and 2
z −1
if z > 1 . (96)

Thus for any test function f on R+∗ one has by (68)

Pfw∫ f u C ( ) 1 −1u d ∗u = PF0 ∫ f (ν )

1
1 −ν
d ∗ν (97)
C

1
with the notations of (69). With f 2 (ν ) = ν 2 f 0 (ν ) we thus get, using (69),

( ) 1 −1u
Pfw∫ f 2 u C d ∗u = PF0 ∫ f 0 f1−1d ∗ν = 2(log 2π + γ ) . (98)
C

We shall now show that,


lim  ∫
f2 u C ∗ ( ) 
d u + log ε  = 0 , (99)
ε →0 1−u C ≥ε 1 − u
 C 

it will then follow that, using (95),

' 1 1
∫ f (u ) 1 − u
C
d ∗u = Pfw∫ f (u )
1− u C
d ∗u . (100)
C

To prove (99) it is enough to investigate the integral,

∫ z ≤1, 1− z ≥ε
[(1 − z )(1 − z )] dz ∧ dz
−1
= j (ε ) (101)

and show that j (ε ) = α log ε + o(1) for ε → 0 . A similar statement then holds for

∫ z ≤1, 1− z −1 ≥ε
[(1 − z )(1 − z )]−1 dz ∧ dz .

One has j (ε ) = ∫ dZ ∧ dZ , where Z = log(1 − z ) and the domain D is contained in the rectangle,
D

 π 
Z = ( x + iy ); log ε ≤ x ≤ log 2,− ≤ y ≤ π / 2 = Rε (102)
 2 

51
and bounded by the curve x = log(2 cos y ) which comes from the equation of the circle z = 1 in
polar coordinates centred at z = 1 . One thus gets,

j (ε ) = 4 ∫
log 2

log ε
( )
Arc cos e x / 2 dx , (103)

when ε → 0 one has j (ε ) ≈ 2π log(1 / ε ) , which is the area of the following rectangle (in the
measure dz ∧ dz ),
{Z = (x + iy ); log ε ≤ x ≤ 0,−π / 2 ≤ y ≤ π / 2}. (104)

One has Rε − 2π log 2 = 2π log(1 / ε ) . When ε → 0 the area of Rε \ D converges to

4∫
−∞
log 2
( )
Arc sin e x / 2 dx = −4∫
π /2

0
log(sin u )du = 2π log 2 , (105)

so that j (ε ) = 2π log(1 / ε ) + o(1) when ε → 0 .

Thus we can assert that with the above choice of basic characters for local fields one has, for any
test function f ,
' 1 1
∫K f (u ) 1 − u d u = Pfw∫ f (u ) 1 − u d u . (106)
∗ ∗

Now, we have the following

Lemma 3.

Let K be a local field, α 0 a normalized character as above and α , α ( x ) = α 0 (λx ) an arbitrary

character of K .

Hence, we obtain that:

' 1 1
∫ f (u ) 1 − u d u = log λ f (1) + Pfw∫ f (u ) 1 − u d u .
∗ ∗
(107)
K

Chapter 4.

On p-adic and adelic strings

4.1 Open and closed p-adic strings.[8]

Let us now discuss the question of the construction of a dynamical theory for open and closed p-
adic strings. It was proposed (Volovich, 1987) to consider p-adic generalization of the Veneziano
string amplitude in two ways, according to two equivalent representations

52
 Γ(a )Γ(b )
1
a −1 b −1
A(a, b ) = ∫ x 1− x dx = . (1)
0
Γ(a + b )

The first way corresponds to an interpretation of the amplitude A(a, b) as a convolution of two
characters and the second one to the p-adic interpolation of the gamma function. Using the first
approach a complex-valued string amplitude over a finite Galois field has been constructed.
Consideration of string amplitudes as a convolution of characters is a very general concept
applicable to characters on number fields, groups and algebras.
Now, we have the string amplitudes of the following form

A(γ a , γ b ) = ∫ γ a ( x )γ b (1 − x )dx , (2)

K

where K is a field F, i.e., K = F, γ a ( x ) is a multiplicative character on K, and dx is a measure on K.

Note that the range of integration in (2) is over the entire field F, and hence this p-adic
generalization is rather one of the Virasoro-Shapiro amplitude

a −1 b −1
A=∫ z 1− z dz , (3)
C

than of the Veneziano amplitude (1), where the integration is over the unit segment on the real axis.
a −1
The equation (3) is just a particular case of (2) for K = C and γ a ( z ) = z . The ordinary Veneziano
amplitude can be rewritten in the following way

a −1 b −1
A=∫ x 1− x θ [0,1] ( x )dx , (4)
R

where θ ( x ) is the characteristic function of the segment [0,1] . In particular, it can be written in
terms of the Heaviside function θ [0,1] ( x ) = θ ( x )θ (1 − x ) . Hence, in order to have a generalization of
the expression (4) on an arbitrary field F one should have on F an analogue of the Heaviside
function or the function sign x.
We have a generalization of the amplitude (4), in the case of an arbitrary locally compact
disconnected field F, in the following form

a −1 b −1
,τ (γ a , γ b ) = ∫ x
AFopen 1− x θτ [0,1]dx (5)
F

where θτ [0,1] ( x ) is a p-adic generalization of the characteristic function of the segment [0,1] on F
( )
related to a quadratic extension F τ . In particular one can take the function θτ [0,1] ( x ) in the form
θτ ( x )θτ (1 − x ) where θτ ( x ) is a p-adic analogue of the Heaviside function.
In the ordinary case there is an important relation between amplitudes of the open and the closed
strings. This relation give a connection on the tree level as follows

closed
Atree (s, t , u ) = sin πt  Atree
open  s t  open  t u 
 ,  Atree  ,  , (6)
8 4 4 4 4

where s, t, u are the Mandelstam variables.

53
Let F in eq. (5) be a non-discrete totally disconnected and locally compact field and define also the
generalized Heaviside function in the form

1 + signτ ω
θτ (ω ) = (7)
2

which is an analogue of the ordinary one.

Now we will consider the amplitude (5) with the characteristic function in one of the following
forms:
θτ [0,1] ( x ) = θτ ( x )θτ (1 − x ) , (8.1)
1
θτ [0,1] ( x ) = (1 + Signτ x ⋅ Signτ (1 − x )) , (8.2)
2
1
θτ [0,1] ( x ) = (Signτ x + Signτ (1 − x )) , (8.3)
2
1
θτ [0,1] ( x ) = (Signτ x − Signτ (− 1) ⋅ Signτ (1 − x )) , τ = ε , (8.4)
2
1
θτ [0,1] ( x ) = (1 − Signτ (− 1) ⋅ Signτ x ⋅ Signτ (1 − x )) . (8.5)
2

The corresponding amplitudes (5) can be calculated with the help of the general formula

Γ(π 1 )Γ(π 2 )
B (π 1 , π 2 ) = , (9)
Γ(π 1π 2 )

which connects the beta function

−1 −1
B (π 1 , π 2 ) = ∫ π 1 ( x ) x π 2 (1 − x )1 − x dx , (10)
F

where π (x ) is a multiplicative character with the gamma function defined by an additive character
χ
−1
Γ(π ) = ∫ χ ( x )π ( x ) x dx . (11)
F

Consider now the string amplitudes, constructed over the p-adic fields Q p and their quadratic
( )
extension Q p τ , from the point of view of the product formulae (6) which relates amplitudes of
closed and open strings in a very simple form. With regard the case τ = ε , the closed string
amplitude defined on the quadratically extended field K = Q p ε , has the form ( )
a −1 b −1 1 − q a−1 1 − q b−1 1 − q c−1
p ( ε ) (a, b, c ) =
AQclosed ∫ x 1− x dx = ⋅ ⋅
1 − q − a 1 − q −b 1 − q −c
, (12)
Qp ( ε)

where q = p 2 . There are no such formulae as simple as (7) for the above constructed open string
amplitudes. However, there exists a formula in the following form

54
 ( ε ) (a, b, c ) = AQp
AQclosed
p
open ,total
(a, b, c )A~Qp (a, b, c ) , (13)

where
a −1 b −1 1 − p a−1 1 − p b−1 1 − p c −1
(a, b, c ) = ∫
,total
AQopen x 1− x dx = ⋅ ⋅ (14)
p
Qp
1 − p − a 1 − p −b 1 − p − c

is a p-adic analogue of the totally crossing symmetric Veneziano amplitude.

Furthermore, the p-adic generalization of the N-point tree amplitude for vector particles in the
bosonic case, can be proposed in the following form

∏(y )
ki k j
A(ζ 1k1 ,..., ζ n k n ) = g n−2 ∫ θ[ ] ( y )F (ζ , k , y )
k n ki k 2 ki
0 , yn −1 ,..., y3 ,1 ⋅ ∏ yi − y j
p i p 1 − yi dyi , (15)
(Q p )
n −3 3≤i < j ≤ n −1 3≤i ≤ n −1

where θ [0, yn−1 ,..., y3,1 ] ( y ) is a p-adic generalization of the characteristic function of the simplex
1 
0 ≤ yn−1 ≤ ... ≤ y4 ≤ y3 ≤ 1 and F (ζ , k , y ) is the part of exp ∑i= j  (ζ iζ j ) / ( yi − y j ) − ki k j / yi − y j 
2

2 
that is multilinear in all the polarization vectors ζ i .

4.2 On adelic strings.[9]

The set of all adeles A may be given in the form

A = U Α (S ) , Α(S ) = R × ∏ Q p × ∏ Z p . (16)
S p∈S p∉S

A has the structure of a topological ring.

We recall that quantum amplitudes defined by means of path integral may be symbolically
presented as
 1 
A(K ) = ∫ A( X )χ  − S [ X ] DX , (17)
 h 

where K and X denote classical momenta and configuration space, respectively. χ (a ) is an

additive character, S [ X ] is a classical action and h is the Planck constant.
Now we consider simple p-adic and adelic bosonic string amplitudes based on the functional
integral (17). The scattering of two real bosonic strings in 26-dimensional space-time at the tree
level can be described in terms of the path integral in 2-dimensional quantum field theory
formalism as follows:

 2πi  4  2πi ( j ) µ 
A∞ (k1 ,..., k 4 ) = g ∞2 ∫ DX exp S 0 [ X ] × ∏ ∫ d 2σ j exp  k µ X (σ j ,τ j ) , (18)
 h  j =1  h 

where DX = DX 0 (σ ,τ )DX 1 (σ ,τ )...DX 25 (σ ,τ ) , d 2σ j = dσ j dτ j and

T
S0 [X ] = −
2∫
d 2σ∂α X µ ∂ α X µ (19)

55
with α = 0,1 and µ = 0,1,...,25 . Using the usual procedure one can obtain the crossing symmetric
Veneziano amplitude

A∞ (k 1,..., k 4 ) = g ∞2 ∫ x ∞1 2 1 − x ∞2 3 dx
kk k k
(20)
R

and similarly the Virasoro-Shapiro one for closed bosonic strings.

As p-adic Veneziano amplitude, it was postuled p-adic analogue of (20), i.e.

Ap (k1 ,..., k 4 ) = g 2p ∫ x p1 2 1 − x p2 3 dx ,
kk k k
(21)
Qp

where only the string world sheet (parametrized by x ) is p-adic. Expressions (20) and (21) are
Gel’fand-Graev beta functions on R and Q p , respectively.
Now we take p-adic analogue of (18), i.e.

 1  4  1 
Ap (k1 ,..., k 4 ) = g 2p ∫ DXχ p  − S 0 [ X ] × ∏ ∫ d 2σ j χ p  − k µ( j ) X µ (σ j ,τ j ) , (22)
 h  j =1  h 

to be p-adic string amplitude, where χ p (u ) = exp(2πi{u}p ) is p-adic additive character and {u}p is
the fractional part of u ∈ Q p . In (22), all space-time coordinates X µ , momenta ki and world sheet
(σ ,τ ) are p-adic.
Evaluation of (22), in analogous way to the real case, leads to

 −1
) .
4
Ap (k1 ,..., k 4 ) = g 2p ∏ ∫ d 2σ j × χ p  ∑ (
ki k j log (σ i − σ j ) + (τ i − τ j )
2 2
(23)
j =1  2hT i< j 

Adelic string amplitude is product of real and all p-adic amplitudes, i.e.

AA (k1 ,..., k 4 ) = A∞ (k1 ,..., k 4 )∏ Ap (k1 ,..., k 4 ) . (24)

p

In the case of the Veneziano amplitude and (σ i ,τ j )∈ Α(S ) × Α(S ) , where Α(S ) is defined in (16),
we have
4
AA (k1 ,..., k 4 ) = g ∞2 ∫ x ∞1 2 1 − x ∞2 3 dx × ∏ g 2p ∏ ∫ d 2σ j × ∏ g 2p .
kk k k
(25)
R
p∈S j =1 p∉S

There is the sense to take adelic coupling constant as

2 2
g A2 = g ∞ ∏ g p = 1 , 0 ≠ g ∈ Q . (26)
p

Hence, it follows that p-adic effects in the adelic Veneziano amplitude induce discreteness of string
momenta and contribute to an effective coupling constant in the form

4
g ef2 = g A2 ∏∏ ∫ d 2σ j ≥ 1 . (26b)
p∈S j =1

56
4.3 Solitonic q-branes of p-adic string theory.[10]

Now we consider the expressions for various amplitudes in ordinary bosonic open string theory,
written as integrals over the boundary of the world sheet which is the real line R. Now replace the
integrals over R by integrals over the p-adic field Q p with appropriate measure, and the norms of
the functions in the integrand by the p-adic norms. Using p-adic analysis, it is possible to compute
N tachyon amplitudes at tree-level for all N ≥ 3 .
This leads to an exact action for the open string tachyon in d dimensional p-adic string theory. This
action is:

1 p2  1 − 12[] 1 
S = ∫ d d xL = 2 ∫ d d
x − φp φ + φ p+1  , (27)
g p −1  2 p +1 

where denotes the d dimensional Laplacian, φ is the tachyon field, g is the open string
coupling constant, and p is an arbitrary prime number.
The equation of motion derived from this action is,

1
− []
p 2
φ = φ p . (28)

The following configuration

φ ( x ) = f (x q+1 ) f (x q+2 )... f (x d −1 ) ≡ F (d −q−1) (x q+1 ,..., x d −1 ) , (29)

with
1
 1 p −1 2 
f (η ) ≡ p 2 ( p −1)
exp − η  , (30)
 2 p ln p 

describes a soliton solution with energy density localised around the hyperplane x q+1 = ... = x d −1 = 0 .
This follows from the identity:
1
− ∂η2
f (η ) = ( f (η )) .
2 p
p (31)

We shall call (29), with f as in (30), the solitonic q-brane solution. Let us denote by
x⊥ = x ( q +1
,..., x d −1
) the coordinates transverse to the brane and by ( )
x|| = x 0 ,..., x q those tangential to
it. The energy density per unit q-volume of this brane, which can be identified as its tension Tq , is
given by
1 p2
Tq = − ∫ d d −q−1 x⊥ L φ = F (d −q−1) (x⊥ ) = 2 (
2gq p + 1
(32) )

where
( d −q −1)/ 4
 p2 −1 
gq = g  . (33)
 2πp
2 p / ( 2 p −1)
ln p 

Hence, we obtain the following equation

57
 1 p2
(
Tq = − ∫ d d −q−1 x⊥ L F (d −q−1) ( x⊥ ) = ) ( d −q −1) / 4 2 p + 1
. (33b)
  p −12
 

2 g  2 p / ( 2 p −1)  
  2πp ln p  

Let us now consider a configuration of the type

φ ( x ) = F (d −q−1) ( x⊥ )ψ (x|| ) , (34)

with F (d −q−1) ( x⊥ ) as defined in (29), (30). For ψ = 1 this describes the solitonic q-brane.
Fluctuations of ψ around 1 denote fluctuations of φ localised on the soliton; thus ψ (x|| ) can be
regarded as one of the fields on its world-volume. We shall call this the tachyon field on the
solitonic q-brane world-volume. Substituting (34) into (28) and using (31) we get

1
− []||
p 2
ψ = ψ p , (35)

where || denotes the (q+1) dimensional Laplacian involving the world-volume coordinates x|| of
the q-brane. The action involving ψ can be obtained by substituting (34) into (27):

1 p2  1 − 12[]|| 1 
S q (ψ ) = S (φ = F (d −q−1) ( x⊥ )ψ (x|| )) =
gq p − 1 ∫
q +1 p +1
2
d x||  − ψ p ψ + ψ , (36)
 2 p + 1 

where g q has been defined in eq.(33).

In conclusion, we shall now show the world-volume action on the Dirichlet q-brane. Let us consider
the situation where we start with the action (27) with g replaced by another coupling constant g ,
and compactify (d – q – 1) directions on circles of radii 1 / 2 . Let u i denote the compact
coordinates and z µ the non-compacts ones, and consider an expansion of the field φ of the form:

C d −q−1 ~ i
φ ( x ) = ψ~ ( z ) + ( (
∑ ξ (z ) 2 cos 2u i + ... .
p i=1
)) (37)

Substituting this into (27), with g replaced by g , we get the action:

1 p 2  2π 
d − q −1
 1 ~ − 1[]z ~  1 ~ i − 12[]z ~ i 1 ~ p−1 ~ i ~ i  
 
g 2 p −1 2  ∫d
q +1
z − ψ p ψ +
2
1 ~ p+1
p +1
ψ − C  ξ p ξ − ψ ξ ξ  + Ο ξ 3 + ... .
~
( )
 2 2 2  
(38)

4.4 Open and closed scalar zeta strings.[11]

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

58
 1 p2  1 −2 
[]
1
Lp = 2  − ϕp ϕ+ ϕ p+1  , (39)
g p −1  2 p +1 

where p is any prime number, = − ∂ t2 + ∇ 2 is the D-dimensional d’Alembertian 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. The eq. (39) take the form:

1  1 1 n+1 
[]
n −1 −
L = ∑ Cn Ln = ∑ 2
Ln = 2 
− φ ∑ n 2
φ +∑ φ . (39b)
n≥1 n≥1 n g  2 n≥1 n ≥1 n + 1 

Recall that the Riemann zeta function is defined as

1 1
ζ (s ) = ∑ s
=∏ −s
, s = σ + iτ , σ > 1 . (40)
n≥1 n p 1− p

Employing usual expansion for the logarithmic function and definition (40) we can rewrite (39b) in
the form
1
L = − 2 [1 / 2φζ ( /2) φ + φ + ln(1 − φ ) ] , (41)
g

where φ < 1 . ζ ( /2) acts as pseudo-differential operator in the following way:

1 ixk  k2 ~ r2
ζ ( /2) φ ( x ) = ∫  2 φ (k )dk , − k = k0 − k > 2 + ε , (42)
 2 2
e ζ −
(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’Alembertian 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
1  k2 ~ φ
(2π )D ∫
ζ ( /2) φ = r eixk ζ  − φ (k )dk = (43)
k02 −k 2 > 2+ε
 2 1−φ

which has an evident solution φ = 0 .

For the case of time dependent spatially homogeneous solutions, we have the following equation of
motion
 − ∂2  1 −ik0t  k 0  ~
2
φ (t )
ζ  t φ (t ) =
 2 
∫
(2π ) k0 > 2 +ε
e ζ 

 2
φ (k0 )dk0 =
1 − φ (t )
. (44)

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

59
 n ( n −1)
1  k2 ~
∫ e ζ  − 2 φ (k )dk = ∑
ixk
ζ ( /2) φ = θ 2 φn , (45)
(2π )D n ≥1

1 ixk  k2 ~  n2 n(n − 1) n (n2−1)−1 n+1 

ζ ( /4) θ = ∫  4 
e ζ  − θ (k )dk = ∑ θ +
2(n + 1)
θ( φ −)1  , (46)
(2π )D n≥1  

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

Chapter 5

On some correlations obtained between some solutions in string theory,

Riemann zeta function and Palumbo-Nardelli Model.

With regard the paper: “Brane Inflation, Solitons and Cosmological Solutions:I”, that dealt various
cosmological solutions for a D3/D7 system directly from M-theory with fluxes and M2-branes, and
the paper: “General brane geometries from scalar potentials: gauged supergravities and accelerating
universes”, that dealt time-dependent configurations describing accelerating universes, we have
obtained interesting connections between some equations concerning cosmological solutions, some
equations concerning the Riemann zeta function and the relationship of Palumbo-Nardelli model.

5.1 Cosmological solutions from the D3/D7 system.[14]

The full action in M-theory will consist of three pieces: a bulk term, S bulk , a quantum correction
term, S quantum , and a membrane source term, S M 2 . The action is then given as the sum of these three
pieces:

S = S bulk + S quantum + S M 2 . (1)

The individual pieces are:

1  1  1
∫d 48  12κ 2 ∫
11
S bulk = 2
x − g R − G 2  − C ∧ G ∧ G , (2)
2κ 

where we have defined G = dC, with C being the usual three form of M-theory, and κ 2 ≡ 8πG N(11) .
This is the bosonic part of the classical eleven-dimensional supergravity action. The leading
quantum correction to the action can be written as:

 1 
S quantum = b1T2 ∫ d 11 x − g  J 0 − E8  − T2 ∫ C ∧ X 8 . (3)
 2 

1/ 3
 2π 2 
The coefficient T2 is the membrane tension. For our case, T2 =  2  , and b1 is a constant
 κ 
number given explicitly as b1 = (2π ) −4 3 −2 2 −13. The M2 brane action is given by:

T2  1 
SM 2 = −
2 ∫ d 3σ − γ γ µν ∂ µ X M ∂ν X N g MN − 1 + ε µνρ ∂ µ X M ∂ν X N ∂ ρ X P C MNP  , (4)
 3 

60
where X M are the embedding coordinates of the membrane. The world-volume metric
γ µν , µ ,ν = 0,1,2 is simply the pull-back of g MN , the space-time metric. The motion of this M2
brane is obviously influenced by the background G-fluxes.

5.2 Classification and stability of cosmological solutions.[14]

The metric that we get in type IIB is of the following generic form:

ds 2 =
f1
α
(− dt 2 + dx12 + dx22 ) + fβ2 dx32 + fγ3 g mn dy m dy n (5)
t t t

where f i = f i ( y ) are some functions of the fourfold coordinates and α , β and γ could be positive
or negative number. For arbitrary f i ( y ) and arbitrary powers of t , the type IIB metric can in
general come from an M-theory metric of the form

2
ds 2 = e 2 Aη µν dx µ dxν + e 2 B g mn dy m dy n + e 2C dz , (6)

with three different warp factors A, B and C , given by:

1 1

1 f f3 1 τ 1 f f3 1 τ 1 f τ2
A = log 1 β2 + log 22 , B = log 3 β2 + log 22 , C = − log β2 + log 2  . (7)
2 α+ 3 τ 2 γ+ 3 τ 3  t τ 
t 3 t 3

To see what the possible choices are for such a background, we need to find the difference B – C .
This is given by:

1 f f τ
B−C = log γ2+ β3 + log 2 . (8)
2 t τ

Since the space and time dependent parts of (8) can be isolated, (8) can only vanish if

τ
f 2 = f 3−1 ⋅ , γ + β = 0 , (9)
τ2

with α and f 1 ( y ) remaining completely arbitrary.

We now study the following interesting case, where α = β = 2 , γ = 0 f 1 = f 2 . The internal six
manifold is time independent. This example would correspond to an exact de-Sitter background,
and therefore this would be an accelerating universe with the three warp factors given by:

2 f 1 1 f  1 f
A= log 21 , B = log f 3 + log 21  , C = − log 21 . (10)
3 t 2 3 t  3 t

We see that the internal fourfold has time dependent warp factors although the type IIB six
dimensional space is completely time independent. Such a background has the advantage that the
four dimensional dynamics that would depend on the internal space will now become time
independent.

61
This case, assumes that the time-dependence has a peculiar form, namely the 6D internal manifold
of the IIB theory is assumed constant, and the non-compact directions correspond to a 4D de-Sitter
space. Using (10), the corresponding 11D metric in the M-theory picture, can then, in principle, be
inserted in the equations of motion that follow from (1). Hence, for the Palumbo-Nardelli model, we
have the following connection:

 R 1 1 
− ∫ d 26 x g − − g µρ g νσ Tr (G µν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 16πG 8 2 

( ) ⇒
∞ 2
1 1 ~ 2 κ
= ∫ 2 ∫ d 10 x(− G ) e − 2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2
1/ 2 2

0 2κ 10  2 g10 
1  1  1
⇒ ∫d 48  12κ 2 ∫
11
2
x − g R − G 2  − C ∧ G ∧ C (11),
2κ 

where the third term is the bosonic part of the classical eleven-dimensional super-gravity action.

5.3 Solution applied to ten dimensional IIB supergravity (uplifted 10-dimensional solution).[14]

This solution can be oxidized on a three sphere S 3 to give a solution to ten dimensional IIB
supergravity. This 10D theory contains a graviton, a scalar field, and the NSNS 3-form among other
fields, and has a ten dimensional action given by

1 1 1 
S10 = ∫ d 10 x g  R − (∂φ ) 2 − e − 2φ H µνλ H µνλ  . (12)
4 2 12 

We have a ten dimensional configuration given by

2 2
3/ 4
 2 2 2 r2  r
5/ 4
 2  Q  
2

ds =  
10  − h ( r ) dt + r dx 0 , 5 + dr 2  +   2 2
dθ + dψ + dϕ +  dψ + cos θdϕ − 5 dt  
r  h( r )  2   5r  

5 r
φ = − log ,
4 2

Q g
H3 = − 6
dr ∧ dt ∧ (dψ + cos θdϕ ) − sin θdθ ∧ dϕ ∧ dψ . (13)
r 2

This uplifted 10-dimensional solution describes NS-5 branes intersecting with fundamental strings
in the time direction.
Now we make the manipulation of the angular variables of the three sphere simpler by introducing
the following left-invariant 1-forms of SU(2):

σ 1 = cosψdθ + sinψ sin θdϕ , σ 2 = sinψdθ − cosψ sin θdϕ , σ 3 = dψ + cos θdϕ , (14)

Q 1
and h3 = σ 3 − dt . (15)
5 r5

62
Next, we perform the following change of variables

4
r 5 ~ 1 ~ 1 ~ 1
= ρ5, t = t , dx 4 = dx 4 , dx5 = dZ , g = 2 g~ , Q = 2 2 7 Q , σ i = ~ σ~i . (16)
2 32 2 2 2 g

It is straightforward to check that the 10-dimensional solution (13) becomes, after these changes

ρ  ~ 2 ~ 2  ~ g~Q 1 ~  
~ 2
1 −1 ~ 2
~ [ ]
ds10 = ρ ds6 + ~ 2 σ 1 + σ 2 +  σ 3 −
2

g  4
d t   + ρdZ 2 ,
2
  4 2 ρ  
φ = − ln ρ ,
~
1 ~ ~ ~ Q ~ ~
H 3 = − ~ 2 σ 1 ∧ σ 2 ∧ h3 + d t ∧ d ρ ∧ h3 , (17)
g 2 g~ρ 5

where we define

2
~ ~ ρ
d~
s 62 = − h ( ρ )d t 2 + ~ dρ 2 + ρ 2 d~
x02, 4 (18)
h (ρ )

and, after re-scaling M,

~ ~
~ 2 M g~ 2 2 Q 2 1
h =− 2 + ρ + . (19)
ρ 32 8 ρ6

We now transform the solution from the Einstein to the string frame. This leads to

1  ~2 ~2  ~ g~Q 1 ~  
~ 2
1 −2 ~ 2
2
10 [ ]
ds = ρ ds6 + ~ 2 σ 1 + σ 2 +  σ 3 −
g  4
d t   + dZ 2 ,
2
  4 2 ρ  

φ = −2 ln ρ ,

H3 = H3 . (20)

We have a solution to 10-dimensional IIB supergravity with a nontrivial NSNS field. If we perform
an S-duality transformation to this solution we again obtain a solution to type-IIB theory but with a
nontrivial RR 3-form, F3 . The S-duality transformation acts only on the metric and on the dilaton,
leaving invariant the three form. In this way we are led to the following configuration, which is S-
dual to the one derived above

1 ~2 ρ 2  ~2 ~2  ~ g~Q 1 ~  
~ 2
2
10 [ ]
ds = ds6 + ~ 2 σ 1 + σ 2 +  σ 3 −
g  4
d t   + ρ 2 dZ 2 ,
2
  4 2 ρ  

φ = 2 ln ρ ,

63
 F3 = H 3 . (21)

With regard the T-duality, in the string frame we have

1 2 r2  2 gQ 1  
2

2
10 [ ]
ds = ds 6 + 2 σ 1 + σ 2 +  σ 3 −
2
4
dt   + r − 2 dZ 2 . (22)
2 g   4 2 r  

This gives a solution to IIA supergravity with excited RR 4-form, C 4 . We proceed by performing a
T-duality transformation, leading to a solution of IIB theory with nontrivial RR 3-form, C 3 . The
complete solution then becomes

1 2 r2  2 gQ 1  
2

2
10 [ ]
ds = ds 6 + 2 σ 1 + σ 2 +  σ 3 −
2
4
dt   + r 2 dZ 2 ,
2 g   4 2r  

φ = 2 ln r

1 Q 1
C3 = − 2
σ 1 ∧ σ 2 ∧ h3 − 5
dt ∧ dr ∧ h3 . (23)
g 2g r

We are led in this way to precisely the same 10D solution as we found earlier [see formula (21)].
With regard the Palumbo-Nardelli model, we have the following connection with the eq. (12):

 R 1 1 
− ∫ d 26 x g − − g µρ g νσ Tr (G µν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 16πG 8 2 

( ) →
∞ 2
1 1 ~ 2 κ
= ∫ 2 ∫ d 10 x(− G ) e − 2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2
1/ 2 2

0 2κ 10  2 g10 
1 1 1 
→ ∫ d 10 x g  R − (∂φ ) − e − 2φ H µνλ H µνλ  . (24)
2

4 2 12 

5.4 Connections with some equations concerning the Riemann zeta function.[14]

We have obtained interesting connections between some cosmological solutions of a D3/D7 system,
some solutions concerning ten dimensional IIB supergravity and some equations concerning the
Riemann zeta function, specifying the Goldston-Montgomery theorem.
In the chapter “Goldbach’s numbers in short intervals” of Languasco’s paper “The Goldbach’s
conjecture”, is described the Goldston-Montgomery theorem.

THEOREM 1

Assume the Riemann hypothesis. We have the following implications: (1) If 0 < B1 ≤ B2 ≤ 1 and
1 X B1
F (X ,T ) ≈ T log T uniformly for 3
≤ T ≤ X B2 log 3 X , then
2π log X

64
 X
1 1
∫ (ψ (1 + δ )x ) −ψ (x ) − δ (x )
2
dx ≈ δX 2 log , (25)
1
2 δ

1 1
uniformly for B2
≤δ ≤ B .
X X 1
X
1 1
∫ (ψ ((1 + δ )x ) −ψ (x ) − δx ) dx ≈ 2 δX
2 2
(2) If 1 < A1 ≤ A2 < ∞ and log uniformly for
1
δ
1 1 1
1 / A1 3
≤ T ≤ 1 / A2 log 3 X , then F (X ,T ) ≈ T log T uniformly for
X log X X 2π

T A1 ≤ X ≤ T A2 .

Now, for show this theorem, we must to obtain some preliminary results .

Preliminaries Lemma. (Goldston-Montgomery)

Lemma 1.

+∞
We have f ( y ) ≥ 0 I (Y ) = ∫ e f (Y + y )dy = 1 + ε (Y ) . If R(y) is a Riemann-
−2 y
∀y ∈ R and let
−∞
integrable function, we have:

b
b 
∫a  (1 + ε ' ( y )) .
R ( y ) f (Y + y )dy = ∫ R ( y )dy 
a 

Furthermore, fixed R, ε ' (Y ) is little if ε ( y ) is uniformly small for Y + a − 1 ≤ y ≤ Y + b + 1 .

Lemma 2.

Let f (t ) ≥ 0 a continuous function defined on [0,+∞ ) such that f (t ) << log 2 (t + 2 ) .

If
T
J (T ) = ∫ f (t )dt = (1 + ε (T ))T log T ,
0
then
∞ 2
 sin ku  π  1
∫0  u  f (u )du =  2 + ε ' (k )k log k ,

with ε ' (k ) small for k → 0 + if ε (T ) is uniformly small for

1 1
2
≤ T ≤ log 2 k .
k log k k

Lemma 3.

65
Let f (t ) ≥ 0 a continuous function defined on [0,+∞ ) such that f (t ) << log 2 (t + 2) . If

∞ 2
 sin ku  π  1
I (k ) = ∫   f (u )du =  + ε ' (k ) k log , (26) then
0
u  2  k

T
J (T ) = ∫ f (t )dt = (1 + ε ' )T log T , (27)
0

1 1
with ε ' small if ε (k ) ≤ ε uniformly for ≤ k ≤ log 2 T .
T log T T

Lemma 4.

4 X i (γ − γ ')
Let F ( X , T ) := ∑
0<γ ,γ '<T 4 + (γ − γ ')
2
. Then (i) F ( X , T ) ≥ 0 ; (ii) F ( X , T ) = F (1 / X , T ) ; (iii) If

The Riemann hypothesis is preserved, then we have

 1  1  log log T 
F ( X , T ) = T  2 log 2 T + log X  + O 

X  2π  log T 

uniformly for 1 ≤ X ≤ T .

Lemma 5.

Let δ ∈ (0,1] and a (s ) =

(1 + δ )s − 1 . If c(γ ) ≤ 1 ∀y we have that
s

2 2
c(γ ) c(γ )
+∞ +∞
 2 1 
∫ a(it ) ∑γ 1 + (t − γ )2 dt = ∫ γ∑≤Z a(1 / 2 + iγ )1 + (t − γ )2 dt + O δ 2 log3 δ  + O Z log3 Z 
2

−∞ −∞

1
for Z > .
δ

For to show the Theorem 1, there are two parts. We go to prove (1).
We define
T 2
X iγ
J ( X , T ) = 4∫ ∑ dt .
0 γ 1 + (t − γ )2

Montgomery has proved that (

J ( X , T ) = 2πF ( X , T ) + O log 3 T )
and thence the hypothesis
1 1
F (X ,T ) ≈ T log T is equal to J ( X , T ) = (1 + o(1))T log T . Putting k = log(1 + δ ) , we have
2π 2

66
 2
 sin kt 
a (it )
2
= 4  .
 t 

For the Lemma 2, we obtain that

∞ 2
X iγ π  1 π  1
∫ a(it ) ∑γ 1 + (t − γ )2 dt =  2 + o(1)k log k =  4 + o(1)δ log δ
2

1 3 1
for ≤T ≤ log 2 .
2 1 δ δ
δ log
δ

For the Lemma 5 and the parity of the integrand, we have that

2
+∞
X iγ π  1
∫−∞ γ∑≤Z 1 + (t − γ )2 dt =  2 + o(1)δ log δ
a ( ρ ) (a)

1 1
if Z ≥ log 3 .
δ δ

X iγ
From the S (t ) = ∑ a (ρ )
γ ≤Z 1 + (t − γ )
2
we note that the Fourier’s transformed verify that

Sˆ (u ) = π ∑ a (ρ ) X iγ e(− γu )e
−2π u
.
γ ≤Z

From the Plancherel identity, we have that

2
+∞
2  1
∫ γ∑ a(ρ )X e(− γu ) du =  + o(1)δ log .
iγ −4π u
e
−∞ ≤Z π  δ

For the substitution Y = log X , − 2πu = y we obtain

2
+∞
1
∫ γ∑ a(ρ )e
iγ (Y + y )
dy = (1 + o(1))δ log
−2 y
e . (b)
−∞ ≤Z δ

Using the Lemma 1 with R( y ) = e 2 y if 0 ≤ y ≤ log 2 and R( y ) = 0 otherwise, and putting

Y+y
x=e we have that
2
2X
3  1
∫ γ∑ a(ρ )x dx =  + o(1)δX 2 log .
ρ

X ≤Z 2  δ

Substituting X with X 2 − j , summarizing on j, 1 ≤ j ≤ K , and using the explicit formula for ψ ( x )

with Z = X log 3 X we obtain

67
 X
1 1
∫ (ψ ((1 + δ )x ) −ψ (x ) − δx ) dx = 2 (1 − 2 )
+ o(1) δX 2 log
2 −2 K
.
X 2− K
δ

Furthermore, we put K = [log log X ] and we utilize, for the interval 1 ≤ x ≤ X 2 − K , the estimate of
Lemma 4 (placing X 2 − K for X ). Thus, we obtain (1).
Now, we prove (2).
We fix an real number X 1 . Making an integration for parts between X 1 and X 2 = X 1 log 2 / 3 X 1 we
obtain, remembering that for hypothesis we have
X
1 2 1
∫1 (ψ ((1 + δ )x ) −ψ (x ) − δx ) dx ≈ 2 δX log δ ,
2

X2
1  1
∫ (ψ ((1 + δ )x ) −ψ (x ) − δx ) x dx =  + o(1)δX 1−2 log .
2 −4
that (c)
X1 2  δ

Utilizing the estimate, valid under the Riemann hypothesis

X
2
∫ (ψ ((1 + δ )x ) −ψ (x ) − δx ) dx << δX
2 2
log 2 ,
1
δ

we obtain analogously as before that

∞
1  1
∫ (ψ ((1 + δ )x ) −ψ (x ) − δx ) x
2 −4
dx << δX 2−2 log 2 = o δX 1−2 log  . (d)
X2
δ  δ

Now, summarizing (c) and (d) and multiplying the sum for X 12 we obtain

∞
 x 2 X 12  1
∫1  X 12 , x 2 (ψ ((1 + δ )x ) −ψ (x ) − δx ) x dx = (1 + o(1))δ log δ .
2 −2
min

Putting X 1 = X , Y = log X , x = eY + y and using the explicit formula for ψ ( x ) with Z = X log 3 X ,
we obtain the equation (b).
2 f
Now, we take the equation (10) and precisely A = log 21 . We note that from the equation (27) for
3 t
T
2 2
ε ' = − and T = 2, we have J (T ) = ∫ f (t )dt = (1 + ε ' )T log T = log 2 . This result is related to
3 0
3
2 f f
A= log 21 putting 21 = 2 , hence with the Lemma 3 of Goldston-Montgomery theorem. Then,
3 t t
we have the following interesting relation

T
2 f
A= log 21 ⇒ ∫ f (t )dt = (1 + ε ' )T log T , (28)
3 t 0

68
hence the connection between the cosmological solution and the equation related to the Riemann
zeta function.
5 r
Now, we take the equations (13) e (21) and precisely φ = − log and φ = 2 ln ρ . We note that
4 2
3
from the equation (27) for ε '= and T = 1/2 , we have
2
T
5 1
J (T ) = ∫ f (t )dt = (1 + ε ' )T log T = log .
0
4 2

T
1
Furthermore, for ε ' = 3 and T = 1/2 , we have J (T ) = ∫ f (t )dt = (1 + ε ' )T log T = 2 log .
0
2

5 r
These results are related to φ = − log putting r = 1 and to φ = 2 ln ρ putting ρ = 1 / 2 , hence
4 2
with the Lemma 3 of Goldston-Montgomery theorem. Then, we have the following interesting
relations:

T T
5 r
φ = − log ⇒ − ∫ f (t )dt = −[(1 + ε ')T log T ] , (29a) φ = 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
4 2 0 0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  (29b)
4 2 12 

hence the connection between the 10-dimensional solutions (12) and some equations related to the
Riemann zeta function.
From this the possible connection between cosmological solutions concerning string theory and
some mathematical sectors concerning the zeta function, whose the Goldston-Montgomery
Theorem and the related Goldbach’s Conjecture.

5.5 The P-N Model (Palumbo-Nardelli model) and the Ramanujan identities.[15]

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:

 R 1 1 
− ∫ d 26 x g − − g µρ g νσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  =
 16πG 8 2 
∞
 2 
( )
2
1 1 ~ 2 κ
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  ,
1/ 2
(30)
0
2κ 10  2 g10 

A general relationship that links bosonic and fermionic strings acting in all natural systems.
It is well-known that the series of Fibonacci’s numbers exhibits a fractal character, where the forms
5 −1
repeat their similarity starting from the reduction factor 1 / φ = 0,618033 = (Peitgen et al.
2
1986). Such a factor appears also in the famous fractal Ramanujan identity (Hardy 1927):

69
 5 −1 5
0,618033 = 1 / φ = = R(q) + , (31)
2 3+ 5  1 q f 5 (−t ) dt 
1+
2
exp ∫ 1/ 5

4/5 
 5 0 f ( −t ) t 

 
 
3  5 ,
and π = 2Φ − R(q) + (32)
20  3+ 5  1 q f 5 (−t ) dt  


1+
2
exp ∫ 1/ 5
 5 0 f (−t ) t  

4/5  

5 +1
where Φ= .
2

Furthermore, we remember that π arises also from the following identity:

π=
12
log 
(
 2 + 5 3 + 13  )(
 , (32a) and π =
)24 
log 
 10 + 11 2 

 10 + 7 2  
+    . (32b)
130 2 142   4   4 
      

The introduction of (31) and (32) in (30) provides:






 R 1 1
− ∫ d 26 x g − ⋅ − g µρ g νσ Tr (Gµν Gρσ ) f (φ ) +
16G   8
  
 3  5 
 2Φ −  R(q ) + 
 20  3+ 5  1 q f (−t ) dt  
5

  1 +
2
exp ∫
 5 0 f −t
1/ 5
(

t 4 / 5  )
 
 
 
1 µν ∞ R 3 5 ⋅
− g ∂ µ φ∂ν φ ] = ∫ 2 ⋅ 2Φ − R (q ) +
2 0 κ 20  3 + 5  1 q f ( −t ) dt  
5
exp ∫ 
11
 1+ 1/ 5 4/5  
 2  5 0 f (−t ) t  
1 ~ 2 κ112
∫ d x(− G )
10 1/ 2
e −2Φ
[ µ
R + 4∂ µ Φ ∂ Φ − H 3 −
2  
Trν
 
3  5  2
2Φ −  R ( q ) + 2 Rg10
20  3+ 5  1 q f (−t ) dt  
5



1+
2
exp
 5
∫ 
0 f ( −t 1 / 5 ) t 4 / 5  

( F ) ],
2
2
(33)

which is the translation of (30) in the terms of the Theory of the Numbers, specifically the possible
connection between the Ramanujan identity and the relationship concerning the Palumbo-Nardelli
model.
70
5.6 Interactions between intersecting D-branes.[12]

Let us consider two Dp-branes in type II string theory, intersecting at n angles inside the ten-
dimensional space.
The interaction between the branes can be computed from the exchange of massless closed string
modes. This can be computed from the one-loop vacuum amplitude for the open strings stretched
between the two Dp-branes, that is given by

dt
Α = 2∫ Tre −tH , (34)
2t

where H is the open string Hamiltonian. For two Dp-branes making n angles in ten dimensions
this amplitudes can be computed to give

tY 2 p −3
dt
) (− iLη (it ) (8πα ' t ) ) (Z
− 2
Α = Vp ∫
0
∞

t
(
exp 2π α ' 8π 2α ' t
−
2
−3 −1 / 2 4 − n
NS − ZR ), (35)

with
n Θ3 (i∆θ j t it ) nΘ (i∆θ t it )
Z NS = (Θ3 (0 it )) − (Θ 4 (0 it ))
4− n 4− n
∏ Θ (i∆θ t it )j =1
∏ Θ (i∆θ t it ) ,
j =1
4 j

1 j 1 j

Θ (i∆θ t it )n
= (Θ (0 it )) ∏
4 −n 2 j
ZR , (36)
Θ (i∆θ t it )
2
j =1 1 j

being the contributions coming from the NS and R sectors. Thence, the eq. (35) can be rewritten
also

tY 2
p −3
dt
) ( )
− 2
Α = Vp ∫
0 t
∞ −
(
exp 2π α ' 8π 2α ' t 2 − iLη (it ) (8πα ' t )
−3 −1 / 2 4− n

(Θ3 (0 it ))4−n ∏ Θ3 (i∆θ j t it ) − (Θ 4 (0 it ))4−n ∏ Θ4 (i∆θ j t it ) ] − [ (Θ2 (0 it ))4−n ∏ Θ2 (i∆θ j t it ) ] . (36b)

n n n
[
j =1 Θ1 (i∆θ j t it ) j =1 Θ1 (i∆θ j t it ) j =1 Θ1 (i∆θ j t it )

Also in (36) Θi are the usual Jacobi functions and η is the Dedekin function. Furthermore, in (35)
by Y we mean the distance between both branes, Y = Σ k Yk2 where k labels the dimensions in
which the branes are separated and Yk the distance between both branes along the k direction.
Now we take the small t limit of (35), that is, the large distance limit (Y >> ls ) . This is the right
limit that takes into account the contributions coming from the massless closed strings exchanged
between the branes.
Using the well known modular properties of the Θ and η functions we obtain, in the t → 0 limit,
that the amplitude is just given by

V p L4−n F (∆θ j ) −
p + n −5
−
tY 2

Α(Y , ∆θ j ) = ( p +1−n ) / 2 ∫t 2
exp 2π 2α '
dt , (37)
2 p −2
(2π α ') 2

71
where the function F contains the dependence on the relative angles between the branes, and is
extracted from the small t limit of (36). The exact form of this function is given by

(4 − n ) + Σ nj+1 cos 2∆θ j − 4Π nj=1 cos ∆θ j

F (∆θ j ) = . (38)
2Π nj=1 sin ∆θ j

Hence, the eq. (37) can be rewritten also

V p L4−n (4 − n ) + Σ nj+1 cos 2∆θ j − 4Π nj=1 cos ∆θ j −

p + n −5
−
tY 2

Α(Y , ∆θ j ) = ( p +1−n ) / 2 ∫t 2
exp 2π 2α '
dt . (38b)
2 p −2
(2π α ')
2 2Π n
j =1 sin ∆θ j

The interaction potential between the branes can then be calculated by performing the integral (37).
This integral is just given in terms of the Euler Γ -function, so the potential has the following form

V p L4−n F (∆θ j )  7 − p − n  ( p+n−7 )

V (Y , ∆θ j ) = − Γ Y . (39)
2 p−2 2π 2α '
p −3
(  2  )
Note that for p + n = 7 this expression is not valid as Γ(0 ) is not a well defined function. In fact in
that case the integral (37) is divergent, so we need to introduce a lower cutoff to perform it. If we
denote by Λ c the cutoff, the integral becomes

V p Lp−3 F (∆θ j ) Y
V (Y , ∆θ j ) = ln . (40)
(4π α ') 2 p −3
Λc

When dealing with compact spaces the expression (37) is modified in the following way

V p L4−n F (∆θ j ) ∞ −
p + n −5
−
tΣ k (Yk + 2πωk R )2

Α(Y , ∆θ j ) = ∑∫ t 2
exp 2π 2α '
dt , (41)
(
2 p−2 2π 2α ' )( p +1− n ) / 2
ωk ∈Z
0

where ωk represents the winding modes of the strings on the directions transverse to the branes.
That means that the summation over k in (41) has only one term in the D6-brane case and it will be
Y9 = x9(1) − x9(2 ) . In the D5-brane case we will have two terms: Y8 = x8(1) − x8(2 ) and Y9 = x9(1) − x9(2 ) .

Also in both cases we will denote Y = Σ k Yk2 . Nevertheless, if the distance between the branes is
small compared with the compactification radii (Y << (2πR )) , the winding modes would be too
massive and then will not contribute to the low energy regime. That is, it will cost a lot of energy to
the strings to wind around the compact space. If we translate this assumption to (41), the dominant
mode will be the zero mode, and the potential can be written as in (39), (40), taking into account
that we focus on the case where the number of angles is n = 2 . In this case the potential, when
normalised over the non-compact directions, for branes of different dimensions is just given by

(2πR )( p−5) F (∆θ j )  5 − p  ( p −5 ) F (∆θ j ) Y

VDp (Y , ∆θ j ) = − Γ Y , (42) VD 5 (Y , ∆θ j ) = ln , (43)
(
2 p−2 2π 2α ' ) p −3
 2  (4π α ')
2 2
Λc

72
where the (2πR )
p −5
factor arises from the dimensions in which the branes become parallel on the
compact dimensions. Furthermore, remember that R denotes the radius of the torus.
Now we note that the eq. (37) can be rewritten substituting to π the corresponding Ramanujan’s
identity (32). Hence, we obtain

V p L4−n F (∆θ j ) p + n −5
−
tY 2

Α(Y , ∆θ j ) =
−

  
∫ t 2
exp 2π 2α '
dt
  
( p +1−n ) / 2  3  5  
2 (2α ')
p −2
2Φ −  R(q ) +
 20 3+ 5  1 q f (− t ) dt   
5

 

1+
2
exp
 5 ∫0 f − t1/ 5 t 4 / 5   
( )
  
(43a)
With regard the eq. (40), we note that can be related with the expression (29b) concerning the
lemma 3 of Goldston-Montgomery Theorem and with the Palumbo-Nardelli Model. Hence, we can
write the following interesting connections:

V p Lp−3 F (∆θ j ) Y
T
ln = 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
(4π α ')
2 p −3
Λc 0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 

1 / 2 −2 Φ 
( ) =
∞
1 1 ~ 2 κ102
⇒∫ 2 ∫
( )
10 µ 2
d x − G e  R + 4∂ µ Φ ∂ Φ − H 3 − 2 Trν F2
0
2κ10  2 g10 
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (43b)
 16πG 8 2 

5.7 General action and equations of motion for a probe D3-brane moving through a type IIB
supergravity background.[44]

Now we will show the general action and equations of motion for a probe D3-brane moving through
a type IIB supergravity background describing a configuration of branes and fluxes.
We start by specifying the ansatz for the background fields that we consider, and the form of the
brane action. We are interested in compactifications of type IIB theory, in which the metric takes
the following general form (in the Einstein frame)

ds 2 = h −1/ 2η µν dx µ dxν + h1/ 2 g mn dy m dy n . (44)

We now embed a probe D3-brane in this background, with its four infinite dimensions parallel to
the four large dimensions of the background solution. The motion of such a brane is described by
the sum of the Dirac-Born-Infeld (DBI) action and the Wess-Zumino (WZ) action. The DBI action
is given, in the string frame, by

73
 S DBI = −T3 g s−1 ∫ d 4ξe −φ − det (γ ab + Fab ) , (45)

where Fab = Bab + 2πα ' f ab , with B2 the pullback of the 2-form field to the brane and f 2 the world-
volume gauge field. γ ab = g MN ∂ a x M ∂ b x N , is the pullback of the ten-dimensional metric g MN in the
string frame. Finally α ' = l 2s is the string scale and ξ a are the brane world-volume coordinates.
The WZ part is given by
SWZ = qT3 ∫ C4 , (46)
W

where W is the world-volume of the brane and q = 1 for a probe D3-brane and q = −1 for a probe
anti-brane. We are interested in exploring the effect of angular momentum on the motion of the
brane, and therefore assume that there are no gauge fields living in the world-volume of the probe
brane, f ab = 0 . For convenience we take the static gauge, that is, we use the non-compact
coordinates as our brane coordinates: ξ a = x µ =a . Since, in addition, we are interested in
cosmological solutions for branes, we consider the case where the perpendicular positions of the
brane, y m , depend only on time. Thus

γ 00 = g 00 + g mn y& m y& n h1/ 2 = −h −1/ 2 (1 − hv 2 ) (47)

and Bab = 0 . Hence

S DBI = −T3 g s−1 ∫ d 4 xe −3φ 1 − hv 2 , (48)

in the Einstein frame. Thence, summing the DBI and WZ actions, we have the total action for the
probe brane
[
S = −T3 g s−1 ∫ d 4 xh −1 e −3φ 1 − hv 2 − q . ] (49)

This action is valid for arbitrarily high velocities. Furthermore, this equation correspond to the
Born-Infeld action for the D-brane embedded in the 10-dimensional space of type IIB theory.
The functions appearing in the following equations

27πα '2  3( g s M )  η 1  c
2
h(η ) =  s
g N +  ln +  = (1 + b lnη ) , (50)
4η 4  2π  η~ 4  η 4
3( g s M )
2
η
N eff = N + ln , (51)
2π η0

are the solutions of the equations of motion for the IIB theory in 10-dimensions, defining the
background. Thence, putting eqs. (50) and (51) in (49), we can determine the trajectory of the brane
in ten dimensions.
Here, η = η~ determines the UV scale at which the KT throat joins to the Calabi-Yau space. This
solution has a naked singularity at the point where h(η ) = 0 , located at η = η~e −1/ b . In this
0 0

configuration, the supergravity approximation is valid when g s M , g s N >> 1 : in this limit the
curvatures are small, and we keep g s < 1 .
We note that also the eqs. (50) and (51), can be related with the expression (29b) and with the
relationship concerning the Palumbo-Nardelli Model. Hence, we obtain the following connections:

74
 T
c c
h(η ) − = b ln η ⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
η4 η 4
0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 

∞
 2 
⇒∫
1
2κ2 ∫d
10 1/ 2

2
1 ~
x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3
g
Trν F2  =
2
−
κ102
2
( )
0 10   10

 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  , (51a)
 16πG 8 2 

2πN + 3(g s M )
2 T
η
ln ⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
2π η0 0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 


( ) =
∞
1 1 ~ 2 κ102
⇒ ∫ 2 ∫ d x(− G ) e  R + 4∂ µ Φ∂ Φ − H 3 − 2 Trν F2
10 1/ 2 µ−2 Φ 2

0
2κ10  2 g10 
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (51b)
 16πG 8 2 

Furthermore, the eq. (49) is also related with the relationship concerning the Palumbo-Nardelli
Model applied to the D-branes. Hence, we have:

∞
− µ 25 ∫ d ξTr e
26
{ −Φ
[− det(Gab + Bab + 2πα ' Fab )] 1/ 2
}= ∫ −
1
(2πα ')2 g YM
2 ∫d
10
{
xTr [− det (η µν + 2πα ' Fµν )]
1/ 2
}
0
∞
 R 1 1  1
⇒ − ∫ d x g − − g µρ g νσ Tr (G µν G ρσ ) f (φ ) − g µν ∂ µ φ∂ν φ  = ∫ 2 ∫ d 10 x(− G ) e −2 Φ
26 1/ 2

 16πG 8 2  0 2κ 10



R + 4∂ µ Φ ∂ µ
Φ −
1 ~ 2 k102
2
H 3 − 2 Trν F2
g10
2
( ) ⇒ − T g

3
−1
s ∫d
4
[
xh −1 e −3φ 1 − hv 2 − q . (52) ]

75
 Chapter 6

Connections.

Now we take the eq. (20) of Chapter 1. We note that can be related with the Godston-Montgomery
equation, the ten dimensional action (12) and the relationship of Palumbo-Nardelli model (30) of
Chapter 5, hence we have the following connection:

k
 1 d 
T
δ k (u ) := δ k ,Β (u ) =  log f u ,Β (Τ) Τ=0 ⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
 λ 'Eˆ (Τ) dΤ 
 Β  0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 
1 / 2 −2 Φ  2 
∞
⇒∫
1
2 ∫
2κ10
d 10
x (− G ) e  R + 4∂ µ Φ ∂ µ
Φ −
1 ~ 2 κ102
2
( )
H 3 − 2 Trν F2  =
g10
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (1)
 16πG 8 2 

Now we take the eq. (29) of Chapter 1. We note that can be related with the equation regarding the
Palumbo-Nardelli model and with the Ramanujan’s identity concerning π . Hence, we have the
following connections:

 
1 2  1 
fϕ , fϕ =
16π 3
( )
N ∏ 1 −  LN 2, ϕ 2 χˆ LN (1,ψ ) ⇒
 qq∉SN  q 
 ϕ 
∞
1
⇒ ∫ π 2 fϕ , fϕ ⋅ ∫ d x(− G )
1/ 2
10
e −2 Φ ⋅
  1 
0
( )
N 2 ∏ 1 −  LN 2, ϕ 2 χˆ LN (1,ψ )GN
 q N  q 
 1 ~ 2 κ2
⋅  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2
2 g10
( ) =
2

 
 
 
 1  1 µρ νσ
= − ∫ d 26 x g  − R ⋅ π 2 fϕ , fϕ ⋅  − g g Tr (Gµν Gρσ ) f (φ ) +
 
   
  8

1
( )
N 2 ∏ 1 −  LN 2, ϕ 2 χˆ LN (1,ψ )G 
  q N  q  
1 
− g µν ∂ µφ∂ν φ  , (2)
2 

76
  
1 2  1 
fϕ , fϕ =
16π 3
(
N ∏ 1 −  LN 2, ϕ 2 χˆ LN (1,ψ ) ⇒ )
 qq∉SN  q 
 ϕ 
 
  1 
⇒
1 2
3
( )
N ∏ 1 −  LN 2, ϕ 2 χˆ LN (1,ψ ) .
    qq∉SN  q 
    ϕ 
 3  5  
162Φ − R(q ) +
 20  3+ 5  1 q f 5 (− t ) dt   
 

1+
2
exp
 5 0
∫ ( )

f − t 1/ 5 t 4 / 5   

(3)

Now we take the eqs. (8) and (9) and (11) of the Chapter 2. We note that can be related with the
Ramanujan’s modular equation (32b) and the Ramanujan’s identity concerning π (32). Thence, we
have the following connection:

π 142
∆ = ∑τ (n )q n = q∏ 1 − q n ( )
24
⇒

⇒
n≥1 n≥1  10 + 11 2   10 + 7 2  
ln   +  
  4   4 
    
 
 
3  5 ⋅ 142
⇒ 2Φ − R(q ) + . (4)
20  3+ 5  1 q f (−t ) dt   
5
 10 + 11 2   10 + 7 2  


1+
2
exp ∫ 1/ 5
 5 0 f (−t ) t   

4 / 5   ln  

+ 
 


 4   4 


Also for the eqs. (11) and (37), we obtain of the similar connections:

(
L∆ (s ) := ∑ τ (n )n − s = ∏ 1 − τ ( p ) p − s + p11 p −2 s ) −1
⇒
n≥1 p

 
 
3 5 142
⇒ 2Φ −  R ( q ) + ⋅ , (5)
20  3+ 5  1 q f (−t ) dt   
5
 10 + 11 2   10 + 7 2  


1+
2
exp ∫ 1/ 5
 5 0 f (−t ) t   

4 / 5   ln  

+ 
 


 4   4 


77
 1
ϕ (s ) = ∏ ⇒
p 1 − τ ( p ) p −s + p11−2 s
 
 
3  5 ⋅ 142
⇒ 2Φ − R(q) + . (6)
20  3+ 5  1 q f (−t ) dt   
5
 10 + 11 2   10 + 7 2  


1+
2
exp
 5 ∫   ln 
0 f ( −t 1 / 5 ) t 4 / 5 
  


+ 
 


 4   4 


Also with regard the eqs. (101) and (106) of Chapter 2, we note that can be related with the
Ramanujan’s identity concerning π . Thence, we have the following connections:

 r  1 c+i∞ 1 ns
∑ 
k∈N ( n / r − a ) 
1 − (k + a )
n  2πi ∫c−i∞ r s
 = ζ ( s , a )
s (s + 1)
ds ⇒

1 c +i ∞ 1 ns
⇒
 
∫ s
  c −i∞ r
ζ (s , a )
s (s + 1)
ds , (7)
  
 3  5  i
22Φ − R(q ) +
 20  3+ 5  1 q f (− t ) dt   
5

 

1+
2
exp
 5 ∫ )

0 f − t 1/ 5 t 4 / 5  
(   
 

1 α +i∞ ns 1 1
∫
2πi α −i∞
ζ ( s , a )
4 s (s + 1)
s
ds = ε (n, a ) − n − + a ⇒
8 2
1 c +i ∞ 1 ns
⇒
 
∫ s
  c −i∞ r
ζ (s , a )
4 s s (s + 1)
ds =
  
 3  5  i
22Φ − R(q ) +
 20  3+ 5  1 q f (− t ) dt   
5

 

1+
2
exp
 5 0
∫ ( )

f − t 1/ 5 t 4 / 5   

1 1
= ε (n, a ) − n − + a . (8)
8 2

Now we take the eqs. (79), (82), (83), (98) and (105) of Chapter 3. We note that can be related
with the Goldston-Montgomery equation (29b) and with the Palumbo-Nardelli relationship (30) of
chapter 5. Hence, we obtain the following connections:

78
 ( ) ( ( )
A = − ∫ ∗ α π −1 y (log q )dy = − log q ∫ α π −1 y dy − ∫ dy =
R R P
) 1
q
log q ⇒
T
⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T ⇒
0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 
1 / 2 −2 Φ 
( ) =
∞
1 1 ~ 2 κ102
⇒∫ 2 ∫
( )
10 µ 2
d x − G e  R + 4∂ µ Φ ∂ Φ − H 3 − 2 Trν F2
0
2κ10  2 g10 
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  , (9)
 16πG 8 2 

1/ 2 T
u
Pfw∫ ∗ f (u )
0
3
d ∗u = log π + γ ⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
R 1− u 0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
 4 2 12 
∞
 2 
1
2κ10
1/ 2 1 ~ 2 κ2
⇒ ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  =
2 g10
( )
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  , (10)
 16π G 8 2 

( ) d ∗u = log(2π ) + lim ∫ ∗ 1 − f 02t f 04 1 − f 04

( ) ( ) d ∗u − log t  = log 2π + γ − log 2
−1 −1
PF0 ∫ ∗ f 04 × 1 − f 04
R+  t →∞  R+ 
T
⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 
∞
 2 
1
2κ10
1/ 2 1 ~ 2 κ2
⇒ ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  =
2 g10
( )
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  , (11)
 16πG 8 2 

79
 Pfw∫ f 2 u C ( ) 1 −1u d ∗u = PF0 ∫ f 0 f1−1d ∗ν = 2(log 2π + γ ) ⇒
C
T
⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 
∞
 2 
1
2κ10
1/ 2 1 ~ 2 κ2
⇒ ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  =
2 g10
( )
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  , (12)
 16πG 8 2 

4∫
log 2

−∞
(
Arc sin e x / 2 dx = −4 ∫ ) 0
π /2
log(sin u )du = 2π log 2 ⇒
T
⇒ 2 ln ρ ⇒ ∫ f (t )dt = (1 + ε ')T log T , ⇒
0

1 1 1 
⇒ ∫ d 10 x g  R − (∂φ ) 2 − e −2φ H µνλ H µνλ  ⇒
4 2 12 
∞
 2 
1
2κ10
1/ 2 1 ~ 2 κ2
⇒ ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  =
2 g10
( )
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (13)
 16πG 8 2 

Now, we take the eqs. (15), (22), (25) and (27) of Chapter 4. We note that can be related with the
Palumbo-Nardelli relationship. Thence, we have the following connections:

∏(y )
ki k j
A(ζ 1k1 ,..., ζ n k n ) = g n−2 ∫ θ[ ] ( y )F (ζ , k , y )
k n ki k 2 ki
0 , yn −1 ,..., y3 ,1 ⋅ ∏ yi − y j i p 1 − yi dyi ⇒
p
(Q p )n −3 3≤i < j ≤ n −1 3≤i ≤ n −1

 R 1 1 
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  =
 16πG 8 2 
1/ 2 −2 Φ 
( ) ,
∞
1 1 ~ 2 κ102
=∫ 2 ∫
( )
10 µ 2
d x − G e  R + 4∂ µ Φ ∂ Φ − H 3 − 2 Trν F2 (14)
0
2κ10  2 g10 

 1  4  1 
Ap (k1 ,..., k 4 ) = g 2p ∫ DXχ p  − S 0 [ X ] × ∏ ∫ d 2σ j χ p  − k µ( j ) X µ (σ j ,τ j ) ⇒
 h  j =1  h 
 R 1 1 
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  =
 16πG 8 2 
1/ 2 −2 Φ 
( ) ,
∞
1 1 ~ 2 κ102
=∫ 2 ∫
( )
10 µ 2
d x − G e  R + 4∂ µ Φ ∂ Φ − H 3 − 2 Trν F2 (15)
0
2κ10  2 g10 

80
 4
AA (k1 ,..., k 4 ) = g ∞2 ∫ x ∞1 2 1 − x ∞2 3 dx × ∏ g 2p ∏ ∫ d 2σ j × ∏ g 2p ⇒
kk k k
R
p∈S j =1 p∉S

 R 1 1 
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  =
 16πG 8 2 
∞
 2 
( )
2
1 1 ~ 2 κ
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  ,
1/ 2
(16)
0
2κ10  2 g10 

1 p2  1 − 12[] 1 
S = ∫ d xL = 2
d
∫
d
d x − φp φ + φ p+1  ⇒
g p −1  2 p +1 
∞
 2 

2κ10
1 10 µ 1 ~ 2 κ102
⇒ ∫ 2 ∫ d x(− G ) e  R + 4∂ µ Φ∂ Φ − H 3 − 2 Trν F2  =
1/ 2

2
−2 Φ

g10
( )
0  
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (17)
 16πG 8 2 

While, if we take the eqs. (18), (33b), (38), (43) and (46) of Chapter 4, we note that can be related
with the Ramanujan’s identity concerning π and with Palumbo-Nardelli model. Then, we obtain
the following connections:

 2πi  4  2πi ( j ) µ 
A∞ (k1 ,..., k 4 ) = g ∞2 ∫ DX exp S 0 [ X ] × ∏ ∫ d 2σ j exp  k µ X (σ j ,τ j ) ⇒
 h  j =1  h 
 R 1 1 
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  =
 16πG 8 2 

( ) ⇒
∞
1 1 ~ 2 κ2
=∫ 2 ∫
d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ −
1/ 2 2
H 3 − 102 Trν F2
0
2κ10  2 g10 
    
    
  3  5  i 1 S X  ×
⇒ g ∞ ∫ DX exp22Φ −
2
R(q ) +
20  f 5 (− t ) dt    h
0
  3+ 5  1 q 
5 ∫0
 1+ exp 
4/5  
 
   2  ( 1/ 5
)
f − t t     

    
     
4
  3  5   k X (σ ,τ ) ,
 i ( )
× ∏ ∫ d σ j exp22Φ −
2
R(q ) + j µ

  20  3+ 5  1 q f (− t ) dt    h
5 µ j j

exp ∫ 
j =1
   1+
0 f ( −t 1 / 5 ) t 4 / 5   
   2  5    
(18)

1 p2
Tq = − ∫ d d − q −1
x⊥ L F ( (d −q −1)
( x⊥ ) = ) ( d −q −1)/ 4 2 p + 1
⇒
  p −12
 

2 g  2 p / ( 2 p −1)  
  2πp ln p  
⇒ − ∫ d d −q−1 x⊥ L(F (d −q−1) ( x⊥ )) =

81
 1 p2
=
(d −q −1) / 4 2 p + 1
   
   
   
   
   
 p2 −1 
2 g  
 
   


   
   3  5   2 p / (2 p−1)  
  2 2Φ −  R(q ) + p ln p 
   20 3+ 5 1 q f 5 (−t ) dt   
 
5 ∫0
 1 + exp( )
4/5 
    2 1/ 5
f ( −t ) t   
   
⇒
∞
{
− µ 25 ∫ d 26ξTr e −Φ [− det (Gab + Bab + 2πα ' Fab )]
1/ 2
}= ∫ −
(2πα ')
1
2 2 ∫d
10
{
xTr [− det (η µν + 2πα ' Fµν )]
1/ 2
}
0 g YM
∞
 R 1 1  1
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  = ∫ 2 ∫ d 10 x(− G ) e −2Φ ⋅
1/ 2

 16πG 8 2  0 2κ10
 1 ~ 2 k2
⋅  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2
2 g10
2
( ) , (19)
 

1 p 2  2π 
d − q −1
 1 ~ − 1[]z ~  1 ~ i − 12[]z ~ i 1 ~ p−1 ~ i ~ i  
 
g 2 p −1 2  ∫  2
d q +1
z − ψ p 2
ψ +
1 ~ p+1
p +1
ψ − C  ξ p ξ − ψ ξ ξ  + Ο ξ + ... ⇒
~3
( )
 2 2  
d − q −1
    
    
 22Φ − 3  R(q ) + 5   
  20  3+ 5  1 q f (−t ) dt    
5
   1 + exp  ∫  
4/5  
1 p 2    2 (
 5 0 f − t t    
1/ 5
)
⇒ 2   ⋅
g p −1  2 
 
 
 
 
 
 1 − []z1
 1 ~ − []z ~ 1
1
~~ 
⋅ ∫ d q+1 z − ψ~p 2 ψ~ +
1 ~ p+1
p +1
ψ − C  ξ i p 2 ξ i − ψ~ p−1ξ iξ i  + Ο ξ 3 + ... ⇒
~
( )
 2 2 2  

{
⇒ − µ 25 ∫ d 26ξTr e −Φ [− det (Gab + Bab + 2πα ' Fab )] = ∫ −
1/ 2
0
∞
} 1
2 2 ∫
(2πα ') gYM
d 10 xTr [− det (η µν + 2πα ' Fµν )] {
1/ 2
}
∞
 R 1 1  1
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  = ∫ 2 ∫ d 10 x(− G ) e −2Φ ⋅
1/ 2

 16πG 8 2  0 2κ10
 2 
µ 1 ~ 2 k102
⋅  R + 4∂ µ Φ∂ Φ − H 3 − 2 Trν F2  ,
2 g10
( ) (20)
 

82
 1  k2 ~ φ
∫k02 −kr2 >2+ε e ζ  − 2 φ (k )dk = 1 − φ ⇒
ixk
ζ ( /2) φ =
(2π )D
1  k2 ~ φ
⇒
 
D ∫ r
k02 − k 2 >2 +ε
eixk ζ  − φ (k )dk =
1−φ
    2
    
  3  5  
22Φ −  R(q ) +
  20 3+ 5  1 q f (−t ) dt   
5

  

1+
2
exp ∫ 
4/ 5  
 5 0 f (−t ) t   
1/ 5
 

 R 1 1 
⇒ − ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  =
 16πG 8 2 
∞
 2 
1
2κ10
1/ 2 1 ~ 2 κ2
= ∫ 2 ∫ d 10 x(− G ) e −2Φ  R + 4∂ µ Φ∂ µ Φ − H 3 − 102 Trν F2  ,
2 g10
( ) (21)
0  

1 ixk  k2 ~  n2 n(n − 1) n (n2−1)−1 n+1 

ζ ( /4) θ = ∫  4 
e ζ  − θ (k )dk = ∑ θ +
2(n + 1)
θ φ −1  ⇒( )
(2π )D n≥1  

1
⇒ D
    
    
  3  5  
22Φ −  R(q ) +
  20 3+ 5  1 q f (−t ) dt   
5

  

1+
2
exp
 5 ∫0 f (−t1/ 5 ) t 4 / 5   
    
ixk  k2 ~  n2 n(n − 1) n (n2−1)−1 n+1 
∫  4 
e ζ  − θ (k )dk = ∑ θ +
2(n + 1)
θ (
φ −1  ⇒ )
n≥1  
1 / 2 −2 Φ 
( ) =
∞
1 1 ~ 2 κ102
⇒∫ 2 ∫
( )
10 µ 2
d x − G e  R + 4∂ µ Φ ∂ Φ − H 3 − 2 Trν F2
0
2κ10  2 g10 
 R 1 1 
− ∫ d 26 x g − − g µρ gνσ Tr (Gµν Gρσ ) f (φ ) − g µν ∂ µφ∂ν φ  . (22)
 16πG 8 2 

Furthermore, we can see easily that the equations described in the Chapter 5 and 6 can be
connected also among them.

Conclusion

Hence, in conclusion, also for some mathematical sectors concerning the Fermat’s Last Theorem,
can be obtained interesting and new connections with other sectors of Number Theory and String
Theory, principally the p-adic and adelic numbers, the Ramanujan’s modular equations, some
formulae related to the Riemann zeta functions and p-adic and adelic strings.

83
Furthermore, also the fundamental relationship concerning the Palumbo-Nardelli model, a general
relationship that links bosonic string action and superstring action (i.e. bosonic and fermionic
strings acting in all natural systems), can be related with some equations regarding the p-adic
(adelic) string sector.

Acknowledgments

I would like to thank Prof. Branko Dragovich of Institute of Physics of Belgrade (Serbia) for the
important and fundamental advices and references that he has give me and his availability and
friendship with regard me. Furthermore, I would like to thank the Prof. G. Tasinato of Oxford
University for his friendship and availability and the Prof. A. Palumbo whose advices has been
invaluable for me. In conclusion, I would like to thank also F. Di Noto for his important
mathematical contribute and useful discussions with regard the Number Theory (Riemann zeta
function and Fibonacci’s Numbers).

References

[1] Wiles Andrew – “Modular Elliptic Curves and Fermat’s Last Theorem” – Annals of
Mathematics, 141 (1995), 443-551.

[2] Edixhoven Bas, Couveignes Jean-Marc, de Jong Robin, Merkl Franz and Bosman Johan
- “On the Computation of coefficients of a modular form” – arXiv:math.NT/0605244v1 – 9
May 2006.

[3] Hammond F. William – “Fermat’s Last Theorem – After 356 Years” – A Lecture at the
Everyone Seminar – University at Albany, Oct 22, 1993 – Minor revisions: 15.07.04.

[4] Deitmar Anton – “Panorama of zeta functions” – arXiv:math.NT/0210060 v4 – 29 Sep 2005.

[5] Riedel Marko Ragnar – “Applications of the Mellin-Perron Formula in Number Theory”
– University of Toronto – August 1996.

[6] Aref’eva I. Ya., Volovich I. V. – “Quantization of the Riemann Zeta-Function and

Cosmology” - arXiv:hep-th/0701284v1 – 30 Jan 2007.

[7] Connes A. – “Trace formula in noncommutative Geometry and the zeros of the Riemann zeta
Function” – arXiv:math.NT/9811068 v1 – 10 Nov 1998.

[8] Aref’eva I. Ya., Dragovich B. G. – “Open and Closed p-adic Strings and Quadratic Extensions
of Number Fields” – CERN-TH.5076/88.

[9] Dragovich B. – “On Adelc Strings” – arXiv:hep-th/0005200 v1 – 22 May 2000.

84
[10] Ghoshal D., Sen A. – “Tachyon Condensation and Brane Descent Relations in p-adic
String Theory” – arXiv:hep-th/0003278 v1 – 30 Mar 2000.

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

[12] Gomez-Reino M., Zavala C. I. – “Recombination of intersecting D-branes and Cosmological

Inflation” – SISSA/ISAS September 06, 2002.

[13] Easson D., Gregory R., Tasinato G and Zavala I. – “Cycling in the Throat” –
arXiv:hep-th/0701252v1 – 29 Jan 2007.

[14] Palumbo A., Nardelli M. – “The Theory of String: A Candidate for a Generalized Unification
Model” – CNRSOLAR 122JA2006 – 22.11.2006.

[15] Nardelli M., Di Noto F., Tulumello A. – “Sulle possibili relazioni matematiche tra Funzione
zeta di Riemann, Numeri Primi, Serie di Fibonacci, Partizioni e Teoria di Stringa” – CNRSOLAR
113BC2006 – 07.11.2006.

Finito di stampare nel mese di Marzo 2007

presso DI. VI. Service – Via Miranda, 50 – 80131 Napoli
Tutti i diritti riservati

85
86