A note on $\theta_2$

Abstract

It is a classic result of Selberg in the 1950's that $\theta_2 = 2/3$, where $\theta_2$ is the level of distribution of the divisor function in arithmetic progressions (defined more precisely below). Selberg applies this estimate, together with his $\Lambda^2$ sieve, to prove weak forms of the binary Goldbach and twin prime conjectures. In this note, we give an unconditional proof of Selberg's first result with smooth weights for prime moduli, and put forth a hypothesis via subconvexity to bootstrap to level $4/5$. Contingent outcomes of our proposal is an improvement of Selberg's second result on approximations to twin primes and in lowering bounds on gaps between primes.