New Criterion for the Riemann Hypothesis

Frank Vega

To cite this version:

Frank Vega. New Criterion for the Riemann Hypothesis. 2024. �hal-04317138v7�

HAL Id: hal-04317138

https://hal.science/hal-04317138v7
Preprint submitted on 18 Jan 2024

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Article
New Criterion for the Riemann Hypothesis
Frank Vega 1

1 NataSquad, 10 rue de la Paix 75002 Paris, Île-de-France, France; [email protected]

 
Abstract: Let Ψ(n) = n · ∏q|n 1 + 1q denote the Dedekind Ψ function where q | n means the prime
Ψ(n)
q divides n. Define, for n ≥ 3; the ratio R(n) = n·log log n where log is the natural logarithm. Let
Nn = 2 · . . . · qn be the primorial of order n. There are several statements equivalent to the Riemann
hypothesis. We prove if for all prime numbers qn (greater than some threshold), there exists another
prime qn′ > qn such that R( Nn′ ) ≤ R( Nn ), then the Riemann hypothesis is true. In this note, using
our criterion, we show that the Riemann hypothesis is true.

Keywords: Riemann hypothesis; prime numbers; Riemann zeta function; Chebyshev function

MSC: 11M26; 11A41; 11A25

1. Introduction
The Riemann hypothesis is a conjecture that the Riemann zeta function has its zeros
only at the negative even integers and complex numbers with real part 12 . It is considered by
many to be the most important unsolved problem in pure mathematics. The hypothesis was
proposed by Bernhard Riemann (1859). The Riemann hypothesis belongs to the Hilbert’s
eighth problem on David Hilbert’s list of twenty-three unsolved problems. This is one of the
Clay Mathematics Institute’s Millennium Prize Problems. In recent years, there have been
several developments that have brought us closer to a proof of the Riemann hypothesis.
There are many approaches to the Riemann hypothesis based on analytic number theory,
algebraic geometry, non-commutative geometry, etc.
The Riemann zeta function ζ (s) is a function under the domain of complex numbers. It
has zeros at the negative even integers: These are called the trivial zeros. The zeta function
is also zero for other values of s, which are called nontrivial zeros. The Riemann hypothesis
is concerned with the locations of these nontrivial zeros. Bernhard Riemann conjectured
that the real part of every nontrivial zero of the Riemann zeta function is 12 .
The Riemann hypothesis’s importance remains from its deep connection to the dis-
tribution of prime numbers, which are essential in many computational and theoretical
aspects of mathematics. Understanding the distribution of prime numbers is crucial for
developing efficient algorithms and improving our understanding of the fundamental
structure of numbers. Besides, the Riemann hypothesis stands as a testament to the power
and allure of mathematical inquiry. It challenges our understanding of the fundamental
structure of numbers, inspiring mathematicians to push the boundaries of their field and
seek ever deeper insights into the universe of mathematics.
In mathematics, the Chebyshev function θ ( x ) is given by

θ (x) = ∑ log q
q≤ x

with the sum extending over all prime numbers q that are less than or equal to x, where log
is the natural logarithm.

Proposition 1. We have [1, pp. 1]:

θ ( x ) ∼ x as ( x → ∞).

We know the following inequalities:

 2 of 10

1
Proposition 2. For r ≥ 0 and −1 ≤ x < r [2, pp. 1]:

1
(1 + x )r ≤ .
1−r·x

Proposition 3. For x > −1 [2, pp. 1]:

x
≤ log(1 + x ) ≤ x.
x+1

Leonhard Euler studied the following value of the Riemann zeta function (1734) [3].

Proposition 4. We define [3, (1) pp. 1070]:

∞ q2k π2
ζ (2) = ∏ q2 − 1 =
6
,
k =1 k

where qk is the kth prime number (We also use the notation qn to denote the nth prime number). By
definition, we have
∞
1
ζ (2) = ∑ 2 ,
n =1 n
where n denotes a natural number. Leonhard Euler proved in his solution to the Basel problem that
∞ ∞ q2k
1 π2
∑ 2
= ∏ 2
=
6
,
n =1 n k =1 q k − 1

where π ≈ 3.14159 is a well-known constant linked to several areas in mathematics such as number
theory, geometry, etc.

The number γ ≈ 0.57721 is the Euler-Mascheroni constant which is defined as

!
n
1
γ = lim − log n + ∑
n→∞ k
k =1
Z ∞ 
1 1
= − + dx.
1 x ⌊x⌋

Here, ⌊. . .⌋ represents the floor function. Franz Mertens discovered some important results
about the constant B (1874) [4].

Proposition 5. Mertens’s second theorem is

!
1
lim
n→∞
∑ q − log log n − B = 0,
q≤n

where B ≈ 0.26149 is the Meissel-Mertens constant [4].

 
In number theory, Ψ(n) = n · ∏q|n 1 + 1q is called the Dedekind Ψ function, where
q | n means the prime q divides n.

Definition 1. We say that Dedekind(qn ) holds provided that

 
1 eγ
∏ 1 + q ≥ ζ (2) · log θ (qn ).
q≤qn
 3 of 10

A natural number Nn is called a primorial number of order n precisely when,

n
Nn = ∏ qk .
k =1

Ψ(n) eγ
We define R(n) = n·log log n for n ≥ 3. Dedekind(qn ) holds if and only if R( Nn ) ≥ ζ (2)
is
satisfied.

Proposition 6. Unconditionally on Riemann hypothesis, we know that [5, Proposition 3 pp. 3]:

eγ
lim R( Nn ) = .
n→∞ ζ (2)

Proposition 7. For all prime numbers qn > 5 [6, Theorem 1.1 pp. 358]:
 
1
∏ 1 +
q
< eγ · log θ (qn ).
q≤qn

The well-known asymptotic notation Ω was introduced by Godfrey Harold Hardy

and John Edensor Littlewood [7]. In 1916, they also introduced the two symbols Ω R and
Ω L defined as [8]:

f (x)
f ( x ) = Ω R ( g( x )) as x → ∞ if lim sup > 0;
x →∞ g (x)
f (x)
f ( x ) = Ω L ( g( x )) as x → ∞ if lim inf < 0.
x →∞ g ( x )

After that, many mathematicians started using these notations in their works. From the last
century, these notations Ω R and Ω L changed as Ω+ and Ω− , respectively. There is another
notation: f ( x ) = Ω± ( g( x )) (meaning that f ( x ) = Ω+ ( g( x )) and f ( x ) = Ω− ( g( x )) are
both satisfied). Nowadays, the notation f ( x ) = Ω+ ( g( x )) has survived and it is still used
in analytic number theory as:

f ( x ) = Ω+ ( g( x )) if ∃k > 0 ∀ x0 ∃ x > x0 : f ( x ) ≥ k · g( x )

which has the same meaning to the Hardy and Littlewood older notation. For x ≥ 2, the
function f was introduced by Nicolas in his seminal paper as [9, Theorem 3 pp. 376], [10,
(5.5) pp. 111]:  
1
f ( x ) = e · log θ ( x ) · ∏ 1 −
γ
.
q≤ x q

Finally, we have the Nicolas Theorem:

1
Proposition 8. If the Riemann hypothesis is false then there exists a real b with 0 < b < 2 such
that, as x → ∞ [9, Theorem 3 (c) pp. 376], [10, Theorem 5.29 pp. 131]:

log f ( x ) = Ω± ( x −b ).

Putting all together yields a proof for the Riemann hypothesis.

2. Central Lemma
Several analogues of the Riemann hypothesis have already been proved. Many authors
expect (or at least hope) that it is true. Nevertheless, there exist some implications in case
of the Riemann hypothesis could be false. The following is a key Lemma.
 4 of 10

Lemma 1. If the Riemann hypothesis is false, then there exist infinitely many prime numbers qn
such that Dedekind(qn ) fails (i.e. Dedekind(qn ) does not hold).

Proof. The function g is defined as [5, Theorem 4.2 pp. 5]:

1 −1
 
eγ
g( x ) = · log θ ( x ) · ∏ 1 + .
ζ (2) q≤ x q

We claim that Dedekind(qn ) fails whenever there exists some real number x0 ≥ 5 for which
g( x0 ) > 1 or equivalent log g( x0 ) > 0 and qn is the greatest prime number such that
qn ≤ x0 . It was proven the following bound [5, Theorem 4.2 pp. 5]:

2
log g( x ) ≥ log f ( x ) − .
x

By Proposition 8, if the Riemann hypothesis is false, then there is a real number 0 < b < 12
such that there exist infinitely many numbers x for which log f ( x ) = Ω+ ( x −b ). Actually
Nicolas proved that log f ( x ) = Ω± ( x −b ), but we only need to use the notation Ω+ under
the domain of the real numbers. According to the Hardy and Littlewood definition, this
would mean that

∃k > 0, ∀y0 ∈ R, ∃y ∈ R (y > y0 ) : log f (y) ≥ k · y−b .

 √ 
The previous inequality is also log f (y) ≥ k · y−b · y · √1y , but we notice that
 √ 
lim k · y−b · y = ∞
y→∞

for every possible values of k > 0 and 0 < b < 21 . Now, this implies that

1
∀y0 ∈ R, ∃y ∈ R (y > y0 ) : log f (y) ≥ √ .
y

Note that, the value of k is not necessary in the statement above. In this way, if the
Riemann hypothesis is false, then there exist infinitely many wide apart numbers x such
that log f ( x ) ≥ √1x . Since √1x > x20 for x0 ≥ 5, then it would be infinitely many wide apart
0
real numbers x0 such that log g( x0 ) > 0. In addition, if log g( x0 ) > 0 for some real number
x0 ≥ 5, then log g( x0 ) = log g(qn ) where qn is the greatest prime number such that qn ≤ x0 .
The reason is because of the equality of the following terms:
  −1   −1
1 1
∏ 1+
q
= ∏ 1+
q
q ≤ x0 q≤qn

and
θ ( x0 ) = θ ( q n )
according to the definition of the Chebyshev function.

3. Main Insight
This is the main insight.

Lemma 2. The Riemann hypothesis is true whenever for each large enough prime number qn , there
exists another prime qn′ > qn such that

R( Nn′ ) ≤ R( Nn ).
 5 of 10

Proof. By Lemma 1, if the Riemann hypothesis is false and the inequality

R( Nn′ ) ≤ R( Nn )

is satisfied for each large enough prime number qn , then there exists an infinite subsequence
of natural numbers ni such that

R( Nni+1 ) ≤ R( Nni ),

qni+1 > qni and Dedekind(qni ) fails. By Proposition 6, this is a contradiction with the fact
that
eγ
lim inf R( Nn ) = lim R( Nn ) = .
n→∞ n→∞ ζ (2)
By definition of the limit inferior for any positive real number ε, only a finite number of
γ
elements of R( Nn ) are less than ζe(2) − ε. This contradicts the existence of such previous
infinite subsequence and thus, the Riemann hypothesis must be true.

4. Main Theorem
This is the main theorem.

Theorem 1. The Riemann hypothesis is true.

Proof. By Lemma 2, the Riemann hypothesis is true if for all primes qn (greater than some
threshold), the inequality
R( Nn′ ) ≤ R( Nn )
is satisfied for some prime qn′ > qn . That is the same as
   
∏q≤qn′ 1 + 1q ∏q≤qn 1 + 1q
≤
log θ (qn′ ) log θ (qn )

and  
∏q≤qn′ 1 + 1q log θ (qn′ )
  ≤
∏q≤qn 1 + 1q log θ (qn )

which is  
1
log log θ (qn′ ) ≥ log log θ (qn ) + ∑ log 1 +
q
qn <q≤qn ′

after of applying the logarithm to the both sides and distributing the terms. That is
equivalent to  
1
log log θ (qn ) ∑ q n < q ≤ q n′
log 1 + q
1≥ +
log log θ (qn′ ) log log θ (qn′ )
after dividing both sides by log log θ (qn′ ). This is possible because of the prime number qn′
is large enough and thus, the real number log log θ (qn′ ) would be greater than 0. We can
apply the exponentiation to the both sides in order to obtain that

! log log1θ (q )
n′
  
log log θ (qn ) 1
e ≥ exp
log log θ (qn′ )
· ∏ 1+
q
.
qn <q≤qn′

For large enough prime qn′ , we have

1
e = (log θ (qn′ )) log log θ (qn′ )
 6 of 10

1
since e = x log x for x > 0. Hence, it is enough to show that
 
1
log θ (qn′ ) > ∏ 1+ .
qn <q≤q ′ q
n

That is equal to  
1
eγ · log θ (qn′ ) > eγ · ∏ 1+
q
.
qn <q≤qn′

By Proposition 7, we know that

 
1
e · log θ (qn′ ) >
γ
∏ 1+
q
.
q≤qn′

So, we deduce that

  −1
1
1>e · γ
∏ 1+
q
q≤qn

which is trivially true since

  −1 !
1
lim
n→∞
eγ · ∏ 1+
q
= 0.
q≤qn

This is because of   −1
1
(log θ (qn ))−1 > ∏ 1+
q
.
q≤qn

We can check that  

lim eγ · (log qn )−1 = 0
n→∞

is true since
θ (qn ) ∼ qn as (n → ∞)
by Proposition 1. Actually, the point here is the statement
  −1
1
(log θ (qn ))−1 > ∏ 1+
q
q≤qn

should be true for large enough n which is equal to say that R( Nn ) > 1 holds indeed. By
Proposition 6, there exists a value of m0 so that for all natural numbers m ≥ m0

eγ eγ
lim inf R( Nm ) − ϵ = − ϵ < R( Nm ) < + ϵ = lim sup R( Nm ) + ϵ
m→∞ ζ (2) ζ (2) m→∞

for every arbitrary and absolute value ϵ > 0 by definition of limit superior and inferior due
to
lim inf R( Nm ) = lim sup R( Nm ) = lim R( Nm ).
m→∞ m→∞ m→∞

In this way, it should exist some value of n0 so that for all natural numbers n ≥ n0 we
γ
obtain that R( Nn ) > 1 since ζe(2) > 1. We would have
 
log log θ (qn )
1 + ϵ1 = exp
log log θ (qn′ )
 7 of 10

and
! log log1θ (q )
n′

1
e · ( 1 − ϵ2 ) = ∏ 1+
q
.
qn <q≤qn′

We only need to prove that

e ≥ ( 1 + ϵ1 ) · e · ( 1 − ϵ2 )
which is the same as
ϵ1
ϵ2 ≥ .
ϵ1 + 1
In addition, we can see that

! log log1θ (q )
n′

1
1−e −1
· ∏ 1+
q
= ϵ2 .
qn <q≤qn′

We have
! log log1θ (q )
! 1
log log θ (q ′ )
n′
  
1 1 n
∏ 1+
q
= 1+ ∏ 1+
q
−1
qn <q≤qn′ qn <q≤qn′
1
≤    
∏qn <q≤q ′ 1+ 1q −1
1− n
log log θ (qn′ )
log log θ (qn′ )
=  
log log θ (qn′ ) + 1 − ∏qn <q≤qn′ 1 + 1q

by Proposition 2, since there always exists a prime number qn′ such that
  !
1
−1 ≤ ∏ 1 + q − 1 < log log θ (qn′ )
qn <q≤q ′ n

due to qn and qn′ are large enough. We can show the inequality
  !
1
∏ 1 + q − 1 < log log θ (qn′ )
qn <q≤q ′ n

could hold for a large enough prime qn′ as well. Indeed, we are able to show that is equal to
  ! !
1 1 1
∑ log 1 + q − q < − ∑ q + log log log(θ (qn′ ))e
qn <q≤q ′
n qn <q≤q ′ n

after of applying the logarithm and adding the term

!
1
− ∑
qn <q≤q ′ q n

to the both sides. By Proposition 3, we verify that

  !
1 1
0≥ ∑ log 1 +
q
−
q
.
qn <q≤qn ′
 8 of 10

By Proposition 5, if we get any large enough prime number qn′ such that
!
1
log log log(θ (qn′ ))e ≥ ∑ ≈ (log log qn′ − log log qn )
qn <q≤q ′ q n

which is
1
(qn′ ) 1+log log θ (qn′ ) ⪅ qn ,
then this could be quite good for supporting our claim. As a consequence, we obtain that

e−1 · log log θ (qn′ )

1−   ≤ ϵ2 .
log log θ (qn′ ) + 1 − ∏qn <q≤qn′ 1 + 1q

Putting all together, we show that

e−1 · log log θ (qn′ ) ϵ1

1−   ≥ .
log log θ (qn′ ) + 1 − ∏qn <q≤qn′ 1 + 1 ϵ1 + 1
q

That is equal to say that

ϵ1 + 1 e−1 · ϵ1ϵ+1 · log log θ (qn′ )

1
−   ≥1
ϵ1 log log θ (qn′ ) + 1 − ∏qn <q≤qn′ 1 + 1q

and
e−1 · (ϵ1 + 1) · log log θ (qn′ )
1≥  .
log log θ (qn′ ) + 1 − ∏qn <q≤qn′ 1 + 1q

where  
1  
1− ∏ 1+
q
≥ e−1 · (ϵ1 + 1) − 1 · log log θ (qn′ )
qn <q≤qn′

after making a simple distribution of the terms. We know that

 
1 log qn′
1− ∏ 1+ ≥−
qn <q≤q ′ q log qn
n

since
   
log qn 1 1
log 1 + + log log qn′ − log log qn ⪆ ∑ ≥ ∑ log 1 +
log qn′ qn <q≤q ′ q qn <q≤q ′ q
n n

by Propositions 3 and 5 whenever qn and qn′ are large enough. That would be

log qn′  
− ≥ e−1 · (ϵ1 + 1) − 1 · log log θ (qn′ )
log qn

that could be rewritten as

log qn′  
≤ 1 − e−1 · (ϵ1 + 1) · log log θ (qn′ )
log qn

which is
e − 1 · ( ϵ1 + 1 )
log log qn′ − log log qn ≤ − + log log log θ (qn′ )
1 − e − 1 · ( ϵ1 + 1 )
 9 of 10

and
ϵ1 + 1
log log qn′ − log log qn ≤ − + log log log θ (qn′ )
e − ϵ1 − 1
by Proposition 3 after of applying the logarithm to the both sides. However, the inequality

ϵ1 + 1
log log qn′ − log log qn ≤ − + log log log θ (qn′ )
e − ϵ1 − 1

trivially holds under the assumption

 
log log θ (qn )
ϵ1 = exp −1
log log θ (qn′ )

and
1
(qn′ ) 1+log log θ (qn′ ) ⪅ qn ,
where qn and qn′ are large enough.

5. Conclusion
The Riemann hypothesis has far-reaching implications for mathematics, with potential
applications in cryptography, number theory, and even particle physics. Certainly, a proof
of the hypothesis would not only provide a profound insight into the nature of prime
numbers but also open up new avenues of research in various mathematical fields.

Acknowledgment
Many thanks to Patrick Solé, Michel Planat and Yusnier Viera for their support.

References
1. Platt, D.J.; Trudgian, T.S. On the first sign change of θ ( x ) − x. Mathematics of Computation 2016, 85, 1539–1547. https:
//doi.org/10.1090/mcom/3021.
2. Kozma, L. Useful Inequalities. Kozma’s Homepage, Useful inequalities cheat sheet. http://www.lkozma.net/inequalities_cheat_
sheet/ineq.pdf, 2011–2024. Accessed 17 January 2024.
3. Ayoub, R. Euler and the Zeta Function. The American Mathematical Monthly 1974, 81, 1067–1086. https://doi.org/10.2307/2319041.
4. Mertens, F. Ein Beitrag zur analytischen Zahlentheorie. J. reine angew. Math. 1874, 1874, 46–62. https://doi.org/10.1515/crll.1874
.78.46.
5. Solé, P.; Planat, M. Extreme values of the Dedekind Ψ function. Journal of Combinatorics and Number Theory 2011, 3, 33–38.
6. Choie, Y.; Lichiardopol, N.; Moree, P.; Solé, P. On Robin’s criterion for the Riemann hypothesis. Journal de Théorie des Nombres de
Bordeaux 2007, 19, 357–372. https://doi.org/10.5802/jtnb.591.
7. Hardy, G.H.; Littlewood, J.E. Some problems of diophantine approximation: Part II. The trigonometrical series associated with
the elliptic ϑ-functions. Acta mathematica 1914, 37, 193–239.
8. Hardy, G.H.; Littlewood, J.E. Contributions to the theory of the Riemann zeta-function and the theory of the distribution of
primes. Acta Mathematica 1916, 41, 119–196.
9. Nicolas, J.L. Petites valeurs de la fonction d’Euler. Journal of Number Theory 1983, 17, 375–388. https://doi.org/10.1016/0022-314
X(83)90055-0.
10. Broughan, K., Euler’s Totient Function. In Equivalents of the Riemann Hypothesis; Cambridge University Press, 2017; Vol. 1,
Encyclopedia of Mathematics and its Applications, pp. 94–143. https://doi.org/10.1017/9781108178228.007.

Short Biography of Authors

Frank Vega is essentially a Back-End Programmer and Mathematical Hobbyist who graduated in
Computer Science in 2007. In May 2022, The Ramanujan Journal accepted his mathematical article
about the Riemann hypothesis. The article “Robin’s criterion on divisibility” makes several significant
contributions to the field of number theory. It provides a proof of the Robin inequality for a large
class of integers, and it suggests new directions for research in the area of analytic number theory.
 10 of 10

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.