Prime number theorem Yaoxin Zhao January 5, 2016 The prime number theorem describes the distribution of the prime numbers among the positive intergers and allows to estimates it with the logarithm. The theroem was urst proven by Jacques Hadamard and Charles Jean de la Vallee-Poussin in 1896. It states the following: Theorem 1. Let π be the prime-counting function dened as: π(x) = {p ∈ N | p ≤ x, p prime}. Then lim x→∞ π(x) log(x) = 1. x There are a number of ways to prove the theorem. Here we take a look at the analytic prove using the Riemann zeta function. Plan of the proof The rst step is to prove that the prime number theorem is equivalent to the statement ψ(x) ∼ x as x → ∞, where ψ(x) is the Chebyshef's function, ψ(x) = X Λ(n), n≤x and Λ : N → R is the Mangoldt function, ( log(p) : ∃m ∈ N, p prime : n = pm Λ(n) = . 0 : otherwise Consider Z ψ1 (x) = ∞ ψ(t)dt. 1 1 (1) Now we can show that 1 ψ1 (x) ∼ x2 as x → ∞ 2 (2) implies (1) and then show (2) itself and the theorem is proven. To prove (2) we express ψ1 (x)/x2 with the Riemann zeta function by means of a contour integral, ζ 0 (x) − ds, where c > 1. ζ(x) x−i∞ P Λ(n)/ns , which is shown in theorem This is shown by using the identity ζ 0 (s)/ζ(s) = − ∞ n=1 1 ψ1 (x) = 2 x 2πi Z x+i∞ xs−1 s(s + 1) 2. Since ζ 0 (s)/ζ(s) has a rst order pole at s = 1 with residue 1, we get ψ1 (x) 1 − x2 2 With 0 Z x+i∞ 1 2 1 xs−1 ζ (x) 1 1− = − ds, − x 2πi x−i∞ s(s + 1) ζ(x) 1−s 1 h(s) = s(s + 1) where c > 1. 0 ζ (x) 1 − − ζ(x) 1−s we have ψ1 (x) 1 − x2 2 Z x+i∞ 1 2 1 xs−1 h(s)ds 1− = x 2πi x−i∞ Z xc−1 ∞ = h(c + it)eit log(x) dt. 2πi −∞ Now we only need to show that xc−1 x→∞ 2πi Z ∞ lim h(c + it)eit log(x) dt = 0. −∞ (3) Showing that the integrant of (3) converges is easily done by applying the RiemannLebesque lemma, but the term xc−1 still tends to innity for c > 1. If we could put c = 1 the term would vanish and the proof would be complete.R The last part of the proof is to ∞ show that we can replace c with 1 and that the integral −∞ | h(1 + it) | dt converges so riemann-Lebesque is still appliable. 2 Theorem 2. Let ζ be the Riemann zeta function, ∀s ∈ C : Re(s) > 1 : ζ(s) = ∞ X 1 . ns n=1 Then for all s with Re(s) > 1: ∞ X Λ(n) ζ 0 (s) =− . ζ(s) ns n=1 Proof. It is shown by Euler Product that the Riemann zeta function can be written as: Y ζ(s) = p prime Then we have: log(ζ(s)) = − 1 . 1 − p−s X log(1 − p−s ) p prime By taking the derivative we get: X log(p)p−s ζ 0 (s) =− ζ(s) 1 − p−s p prime X 1 =− log(p) −1 1 − p−s p prime X =− log(p) p prime ∞ X p−ms m=1 For p prime we have: ∞ X n:Λ(n)=log(p) X Λ(n) = log(p) p−ms ns m=1 Then by denition of the Mangoldt function: ∞ X Λ(n) n=1 ns = X p prime and the proof is complete. 3 log(p) ∞ X m=1 p−ms