$\mathrm{Zi}_n(x)$ limits to $\pi(x).$ But exactly?

110 Views Asked by At

Consider the prime counting function: $$\mathrm{Zi}(x)=\frac{1}{e}\sum_{k=1}^\infty\frac{(\ln x)^k}{kk!\phi(k)}$$

where $$\phi(k)=\sum_{n=1}^\infty e^{-{n^k}}$$

This page shows how to sum the nontrivial zeros of the Riemann zeta function in Riemann's prime counting function.

What happens when you follow this prescription for $\mathrm{Zi}(x)?$ Do you exactly obtain $\pi(x)?$ Is there some other set of "nontrivial zeros" that would give exactly $\pi(x)?$

$$ \mathrm{Zi}_n(x)=\mathrm{Zi}(x)-\sum_{m=1}^\infty \mathrm{Zi}(x^{-2m})+ \sum_{k=1}^n T_k(x) $$

where $$ T_k(x)= - \mathrm{Zi}(x^{\rho_{k}})-\mathrm{Zi}(x^{\rho_{-k}}) $$

My guess is that in the limit you'd obtain a function very close to the prime counting function $\pi(x)$ but it wouldn't be exact.

Here is a plot of $\pi(x)$ and $\mathrm{Zi}(x):$

enter image description here