Analytic Number Theory Exponent Database

18 Zero free region for the zeta function

This chapter is not yet integrated into the main blueprint.

A zero of the Riemann zeta function is a complex number ρ=β+iγ for which ζ(ρ)=0. The zeta function has a infinite number of zeros of the form ρ=2n for integer n1; these are known as trivial zeros and are well understood. There are also an infinite number zeros inside the “critical strip" 0<z<1, called non-trivial zeros. The locations of the non-trivial zeros have deep consequences for many fields of mathematics. A well-known conjecture regarding the non-trivial zeros is the Riemann hypothesis.

Heuristic 18.1 Riemann hypothesis
#

If ρ is a non-trivial zero of the Riemann zeta function, then ρ=1/2.

This conjecture remains far out of reach. Instead, currently what are known are zero free regions.

Definition 18.2 Zero free region
#

A zero-free region of the Riemann zeta function is a set DC for which ζ(s)0 for all sD.

Lemma 18.3 Basic properties of zero free regions

The following properties hold:

  1. (Symmetry about the real axis) If ζ(σ+it)0 then ζ(σit)0.

  2. (Symmetry about the critical line s=1/2) For 0σ1, if ζ(σ+it)0 then ζ(1σ+it)0.

  3. (Non vanishing for s>1) If s>1 then ζ(s)0.

Proof

In light of Lemma 18.3, for the rest of the chapter we will focus on the quadrant

D{zC:z>1/2,z>0}.

The first zero free region for non-trivial zeros proved was on D={zC:z=1}, and was used to prove the prime number theorem π(x)x/logx as x.

Theorem 18.4 Non-vanishing on the 1-line
#

One has ζ(1+it)0 for any real t.

18.1 Relation to upper bound on zeta in the critical strip

Using estimates of ζ(σ+it) close to the line σ=1, one can extend the zero free region slightly inside the critical strip.

Lemma 18.5 Relation to growth exponents of zeta

Suppose f(t)>0 and 0<g(t)1 are real-valued functions for t0, with f(t) non-decreasing and tending to infinity with t, and g(t) non-increasing. Suppose further that f(t)/g(t)=o(exp(f(t))). If

ζ(σ+it)exp(f(t))(1g(t)σ2,t0)

then ζ(σ+it)0 for

σ1Ag(2t+1)f(2t+1)

where A>0 is an absolute constant.

Proof

The following zero free region of classical type was proved independently by de la Vallée Poussin and Hadamard.

Theorem 18.6 Classical zero free region

One has ζ(σ+it)0 if

σ1Alogt.

for an absolute constant A>0 and t sufficiently large.

Proof

This classical result has been improved in a number of works, most of which make crucial use of non-trivial estimates of certain types of exponential sums.

Theorem 18.7 Littlewood zero free region

One has ζ(σ+it)0 if

σ1Aloglogtlogt

for an absolute constant A>0 and t sufficiently large.

Proof
Theorem 18.8 Chudakov zero free region
#

One has ζ(σ+it)0 if

σ11(logt)3/4+o(1)

for t sufficiently large.

Theorem 18.9 Korobov-Vinogradov zero free region
#

One has ζ(σ+it)0 if

σ1A(logt)2/3(loglogt)1/3

for an absolute constant A>0 and t sufficiently large.

18.2 Relation to the error term in the prime number theorem

Lemma 18.10 Relation to prime number theorem error term

Suppose ζ(σ+it)0 for σ1η(t) where η(t) is a positive and decreasing function. Then

nxΛ(n)xxexp(Aω(x)),(x)

for an absolute constant A>0, where

ω(x):=inft1(η(t)logx+logt).
Proof

Applying Lemma 18.10, one obtains the error term estimates in the prime number theorem given in Table 18.1.

Table 18.1 Error bounds on the prime number theorem. Here A represents an absolute, positive constant, which may be different at each occurrence.

Bound on (ψ(x)x)/x

Associated zero-free region

Reference

exp(A(logx)1/2)

σ1Alogt

Theorem 18.6

exp(A(logxloglogx)1/2)

σ1Aloglogtlogt

Theorem 18.7

exp(A(logx)3/5(loglogx)1/5)

σ1A(logt)2/3(loglogt)1/3

Theorem 18.9