An arithmetical property of powers of Salem numbers (Q2506380)

In [Bull. Lond. Math. Soc. 38, No. 1, 70--80 (2006; Zbl 1164.11025)] the reviewer established a lower bound for the difference \(\Delta(\zeta,\alpha)\) between the largest and the smallest limit points of the sequence of fractional parts \(\{\zeta \alpha^n\},\) \(n=0,1,2,\dots,\) where \(\alpha>1\) is an algebraic number and \(\zeta \neq 0.\) An additional requirement is that \(\zeta \notin {\mathbb Q}(\alpha)\) in case \(\alpha\) is a Salem or a Pisot number. Here, the author considers one of these exceptional cases, namely, the case when \(\alpha\) is a Salem number (say, of length \(L(\alpha)\)) and \(\zeta\) is a nonzero element of the field \({\mathbb Q}(\alpha)\). He proves that \(\Delta(\zeta,\alpha) \geq 1/L(\alpha)\) if \(\alpha-1\) is a unit. If, alternatively, \(\alpha-1\) is not a unit then he shows that for every \(\varepsilon>0\) there is a \(\zeta \in {\mathbb Q}(\alpha)\) such that \(\Delta(\zeta,\alpha)<\varepsilon\). Furthermore, he proves that \(\Delta(\zeta,\alpha)=1\) if \(\zeta\) is an algebraic integer, so that both \(0\) and \(1\) are the limit points of the sequence \(\{\zeta \alpha^n\},\) \(n=0,1,2,\dots.\) For \(\zeta=1,\) an old result of Salem implies that this sequence is everywhere dense in \([0,1]\). In his proof the author uses arguments from linear algebra, diophantine analysis and uniform distribution.