A question on the relation between the complex zeroes of zeta function and the estimate of the error in PNT

46 Views Asked by At

I'm currently working on the following problem from Analytic Number Theory.

Assume that $\psi(x)-x=\mathcal{O}(x^a)$, for some $1/2<a<1$, where $\psi$ is the Chebyshev function.

I would like to show that this implies that $\zeta(s)$ has no zeroes with $\sigma>a$, where $s=\sigma+it$.

I'm not sure how to proceed with the proof of this claim. Any hint/help will be very useful. Thank you.