On the digital representation of integers with bounded prime factors (Q1750369)

The present note considers the \(b\)-adic expansion of integers containing only small prime factors. In particular, the author wants to show that there is no connection between the size of the prime factors and the number of nonzero digits. Let \(S=\{q_1,\ldots,q_s\}\) be a finite, non-empty set of distinct prime numbers. Let \(n\) be a positive integer. Then we may write \(n=q_1^{r_1}\cdots q_s^{r_s}n'\) where \(r_1,\ldots,r_s\) are non-negative integers and \(n'\) is coprime with every \(q_i\). Then we define the \(S\)-part \([n]_S\) of \(n\) by \[[n]_S:=q_1^{r_1}\cdots q_s^{r_s}.\] For a positive integer \(b\) let \(u_j^{(k)}\) be the ordered sequence of positive integers not divisible by \(b\) and whose \(b\)-ary representation has at most \(k\) nonzero digits, \textit{i.e.} \[d_kb^{n_k}+\ldots+d_2b^{n_2}+d_1,\quad n_k>\cdots>n_2>0,\quad d_1\ldots,d_k\in\{0,1,\ldots,b-1\},\quad d_1d_k\ne 0.\] Then the author's first result states that for integers \(b\) and \(k\ge 2\) and \(\varepsilon>0\) \[[u_j^{(k)}]_S<(u_j^{(k)})^\varepsilon,\] for every sufficiently large integer \(j\). The proof is not effective as it uses the Schmidt Subspace Theorem. However, the author shows for the case \(k=3\) a lower bound for the size of the largest prime factor of \(u_j^{(k)}\). A lower bound for the case \(k=2\) follows from a combination of results by \textit{Y. Bugeaud} and \textit{J.-H. Evertse} [Mathematika 63, No. 3, 840--851 (2017; Zbl 1434.11042)] and \textit{C. L. Stewart} [in: Number theory and related fields. In memory of Alf van der Poorten. Based on the proceedings of the international number theory conference, Newcastle, Australia, March 12--16, 2012. New York, NY: Springer. 341--359 (2013; Zbl 1315.11011)].