On multiplicatively independent bases in cyclotomic number fields (Q519981)

For any positive integer \(k\) let \(\Phi_k\) be the \(k\)-th cyclotomic polynomial and \(\zeta_k\) be any of its roots, i.e., any \(k\)-th primitive root of unity. A number \(-m+\zeta_k\), where \(m\in\mathbb{Z}\), is called a base of a numeration system in \(\mathbb{Z}[\zeta_k]\) if for every \(\alpha\in\mathbb{Z}[\zeta_k]\) there is a unique polynomial \(f\in\mathbb{Z}[x]\) having all coefficients in \(\{0,1,\dots,\Phi_k(-m)-1\}\) such that \(\alpha=f(-m+\zeta_k)\). One of the results of the paper states that if \(m\leq0\) then \(-m+\zeta_k\) is not a base of a numeration system in \(\mathbb{Z}[\zeta_k]\). Two algebraic integers \(\alpha\), \(\beta\) are called multiplicatively independent if \(\alpha^r=\beta^s\) with \(r,s\in\mathbb{Z}\) holds only for \(r=s=0\). The main result of the paper is the following theorem: Suppose that \(a>0\), \(k>2\), and \(m\) are integers such that \(-m+\zeta_k\) and \(-(m+a)+\zeta_k\) are multiplicatively dependent and none of them is a root of unity. Let \(S\) be the set of all prime numbers dividing \(\Phi_k(m)\cdot\Phi_k(m+a)\). Then either \(m=-1\), \(a=2\), and \(k=4\), or \(\bullet\) \(S\) is a nonempty subset of the set of all prime numbers dividing \(a\); \(\bullet\) if \(k\not\equiv2\pmod 4\) then for each \(p\in S\) we have \(k\leq p^{\nu_p(a)}\), \(p\nmid k\), and \(\operatorname{ord}_k(p)\leq\nu_p(a)\); \(\bullet\) for both \(x\in\{m,m+a\}\), we have \(\operatorname{ord}_k(p)\mid\nu_p(\Phi_k(x))\) and \(\nu_p(\Phi_k(x))\leq\nu_p(a)\). Here \(\operatorname{ord}_k(p)\) means the order of \(p\) modulo \(k\) and \(\nu_p\) is the \(p\)-adic valuation on \(\mathbb{Z}\). The authors used the above theorem to get the following computational result: if \(1\leq a\leq10^6\), \(m\neq0\), \(m\neq-a\), and \((m,k)\notin\{(1,6),(-1,3),(-a+1,6),(-a-1,3)\}\) then \(-m+\zeta_k\) and \(-(m+a)+\zeta_k\) are multiplicatively independent or \(m=-1\), \(a=2\), and \(k=4\).