The Distribution of Error Terms of Smoothed Summatory Totient Functions

Abstract

We consider the summatory function of the totient function after applications of a suitable smoothing operator and study the limiting behavior of the associated error term. Under several conditional assumptions, we show that the smoothed error term possesses a limiting logarithmic distribution through a framework consolidated by Akbary--Ng--Shahabi. To obtain this result, we prove a truncated version of Perron's inversion formula for arbitrary Riesz typical means. We conclude with a conditional proof that at least two applications of the smoothing operator are necessary and sufficient to bound the growth of the error term by $\sqrt{x}$.