The Zeta Function of Riemann
Mani Tadayon
Bernhard Riemann was not the first mathematician to study the zeta func-
tion which today bears his name. However, his paper, Über die Anzahl der
Primzahle unter einer gegebenen Grösse (On the number of primes less than a
given magnitude), published in 1859, placed the zeta function at the focal point
of mathematical inquiry for generations. This paper will trace the history of
the zeta function up to Riemann’s paper, examine in very limited manner the
Riemann hypothesis, and the significance of Riemann’s zeta function in modern
mathematics. We will focus specifically on the application of the zeta function
to the density and distribution of primes.
The zeta function is commonly defined as
∞
X 1
ζ(s) =
ks
k=1
where s is a complex number. It should be noted that this definition is only one
of several. Other definitions of the zeta function relate it to advanced functions
which are beyond the scope of this paper, including the gamma function, Mellin
transform, Bernoulli numbers, and complex multiple integrals. A cursory glance
at the zeta function as we have defined it establishes its link to integration,
since it is an infinite sum, and any converging value would be an infinite sum of
increasingly infinitesimal values. The importance of the zeta function in varied
branches of mathematics is evidenced by its relation to many special functions.
The zeta function can be used to classify numbers in categories more specific
than prime or not-prime, based on the form of their prime factors. Examples of
this are the use of the zeta function in studying powerful numbers and almost-
primes.
A number is considered k -almost-prime when the sum of the powers of its
prime factors is k. A prime number, whose prime factors all all have exponent
0, would be 0-almost-prime, and a number with only 2 prime factors would
be 2-almost prime. Powerful numbers are those whose prime factors all have
powers greater than or equal to 2.
We will work with one alternative definition of the zeta function, which
relates the zeta function to the Euler product
∞
X 1 Y 1
ζ(s) = =
k
k=1
s
p
1 − p1s
1
2
where the product is taken over all primes p. This equality is also known as the
Euler product formula of the zeta function.
Riemann began his paper by discussing the Euler product. Leonhard Euler
was the first to study the zeta function in detail, and established its relation to
prime number theory. One of Euler’s myriad papers (his work accounted for a
quarter of all mathematics publications in the 18th century), published in 1737,
proved that the following sum diverges:
X1 1 1 1 1 1 1
= + + + + + + ...
p
p 2 3 5 7 11 13
where the sum is over all primes p. This discovery foreshadowed Riemann’s
analysis of the density of primes. It should be noted that Euler considered the
zeta function only for real values of s. The proof for the divergence of the sum
of the reciprocals of primes is actually more difficult than the proof of the Euler
product formula. In fact, the divergence of the sum of the reciprocals of primes
follows directly from the Euler product formula, as does the infinitude of primes.
We proceed to prove that
∞
X 1 Y 1
ζ(s) = =
ks
k=1 p
1 − p1s
This classical proof follows the pattern of a typical analysis of infinite series,
namely subtracting multiples of the series to find a formula for the series. The
novelty lies in the concept of taking a product over all primes. Although it
may be impossible to find a formulaic expression for the terms in this product,
the concept
Q is quite rigorous, and does not require much abstraction. The
notation p f (p) means to take the value of f for every prime number and
Q∞
form a product. One could also use subscript notation n=1 f (pn ) where pj
denotes the jth prime. This notation is closer to standard series notation, but
detracts from the concept of taking the primes as an infinite set, rather than as
a collection of individual values.
We start with
∞
X 1 1 1 1 1
ζ(s) = = 1 + s + s + s + s + ...
ks 2 3 4 5
k=1
and multiply both sides by (1 − 21s )
1 1 1 1 1 1 1 1 1 1
ζ(s) 1 − s = 1 + s + s + s + s + . . . − s 1 + s + s + s + s + . . .
2 2 3 4 5 2 2 3 4 5
1 1 1 1 1 1 1 1 1 1 1
ζ(s) 1 − s = 1+ s − s + s + s − s + s +. . . = 1+ s + s + s + s +. . .
2 2 2 3 4 4 5 3 5 7 9
This eliminates all multiples of 12 in ζ(s)(1 − 21s ) when expanded. Then we
multiply both sides by (1 − 31s )
1 1 1 1 1 1 1 1 1 1 1
ζ(s) 1 − s 1 − s = 1 + s + s + s + s + ... − s 1 + s + s + s + s + ...
2 3 3 5 7 9 3 2 3 4 5
3
1 1 1 1 1 1 1 1 1 1 1 1
ζ(s) 1 − s 1 − s = 1+ s − s + s + s + s − s +. . . = 1+ s + s + s + s +. . .
2 3 3 3 5 7 9 9 5 7 11 13
This eliminates all multiples of 31 in ζ(s)(1 − 21s )(1 − 31s ). If we continue to
multiply both sides by 1 − p1s for all primes p, we will eventually eliminate all
multiples of p1 for all primes p, that is every term except 1. We can write this
as
Y 1
ζ(s) 1− s =1
p
p
or
1 Y 1
ζ(s) = Q 1 =
p (1 − ps ) p
1 − p1s
Now, returning to Euler’s theorem on the divergence of the sum of the re-
ciprocals of all primes, we examine the case of ζ(1), or more precisely, the
right-handed limit
Y 1
lim+ ζ(s) = lim+
s→1 s→1
p
1 − p1s
We note that
1 ps
1 = ps − 1
1 − ps
is greater than 1 for s ≥ 1, therefore the product of such terms would grow
to infinity as the number of terms grew to infinity. If there were only a finite
number of primes, then the product would also be finite, i.e.
∞
Y 1 X 1
lim = lim
s→1+
p
1 − p1s s→1+
k=1
ks
would be a converging sum. However, we know that the harmonic sum
∞
X 1
k
k=1
is divergent. Therefore there must be an infinite number of primes, and thus
the sum of the reciprocals of primes are themselves a divergent series (this last
conclusion requires significantly more elaboration to be rigorous, but we shall
not pursue a complete proof).
This preliminary result is already enough to intimate the importance of the
zeta function to the density of primes. Although at first glance, the zeta function
does not seem to have any relation to the primes, we have shown that it indeed is
a variant of an infinite product of a function over all primes numbers. Although
the function,
Y 1
p
1 − p1s
4
appears to be somewhat exotic and arbitrary, it is balanced by the relative
simplicity of the zeta function itself. Further, the zeta function tells us that
the infinitude of primes and their density are deeply related to the properties
of all numbers. The relation between an infinite sum over all numbers to an
infinite product over all primes is actually what our intuition would tell us if
we consider that any set of integers can be reduced to a smaller set of prime
numbers (since all integers are products of primes). The link between products
of primes and the integers is thus at the center of how our system of numbers
is constructed. The connection between complex analysis and number theory is
an important aspect of the zeta function.
Euler also defined the factorial function for real numbers greater than −1,
which is written as Π(s), or alternatively as Γ(s) = Π(s − 1)
Z ∞
s! = Π(s) = Γ(s + 1) = e−x xs dx
0
We note that s! is defined only for positive integer values of s. Integrating
Γ(s + 1) by parts, we have
Z ∞ Z ∞
−x s
s+1 −x ∞
e x dx = x e 0
+s e−x xs−1 dx
0 0
R∞
We note that 0
e−x xs−1 dx = Γ(s), and we can simplify further
Γ(s + 1) = sΓ(s)
which is the same recursive definition of the factorial function. We must also
show that Γ(1) = 1 (note that we cannot use our recursive definition since the
factorial function is not meaningful for 0, although 0! is commonly assigned a
value of 1).
Z ∞ Z ∞
∞ 1
e−x x0 dx = e−x dx = −e−x 0 = − ∞ − 1 = 1
Γ(1) =
0 0 e
These functions play a critical role in a detailed study of the Riemann zeta
function, requiring some advanced complex analysis, so we shall employ them
2
sparsely. Euler also found values for ζ(2) = π6 and other integer values, as well
as values for ζ(2n) which involves powers of π and functions of the Bernoulli
numbers.
We should note that we have not yet touched upon Riemann’s contribution
to the theory of the zeta function. One of Riemann’s contributions was the
following relation between ζ(s) and ζ(1 − s), known as the functional equation
of the zeta function
s 1−s (1−s)
Π − 1 ζ(s) = Π − 1 π − 2 ζ(1 − s)
2 2
Riemann then went on to analyze many aspects of the zeta function, making
two critical observations. The first was that the zeta function was an asymptote
5
of π(x), which gives the number of primes less than x. This can be expressed
ζ(s)
lim =1
s→∞ π(s)
The distribution of primes over the natural numbers has been surmised as
approximately log x, and empirical results have reinforced this view. The zeta
function has been shown to be a much better estimate for the distribution of
primes than the natural logarithm. It is also widely held that the zeta function
gives the exact distribution of primes over the natural numbers.
The second observation was a conjecture known as the Riemann Hypothesis,
which has yet to be proved (although it is strongly suspected of being true). The
Riemann hypothesis states that all non-trivial zeroes of the zeta function (over
the domain of complex numbers) have real part of 21 . Furthermore, there are an
infinite number of zeroes with real part 12 . A proof to the Riemann Hypothesis
is perhaps the most sought after solution in mathematics today.
Bibliography
Edwards, H.M. , Riemann’s Zeta Function, Academic Press, New York 1977.
Volkovyskii, L.I., et al, translated by J. Berry, A Collection of Problems on
Complex Analysis, Pergamon Press, Oxford 1965
Patterson, S.J., An introduction to the theory of the Riemann Zeta-Function,
Cambridge University Press, Cambridge 1988.
Ivić, Aleksandar, The Riemann Zeta-Function, John Wiley and Sons, New
York 1985.
