Primitive prime divisors in zero orbits of polynomials

From MaRDI portal



Abstract: Let (bn)=(b1,b2,...) be a sequence of integers. A primitive prime divisor of a term bk is a prime which divides bk but does not divide any of the previous terms of the sequence. A zero orbit of a polynomial f(z) is a sequence of integers (cn) where the n-th term is the n-th iterate of f at 0. We consider primitive prime divisors of zero orbits of polynomials. In this note, we show that for integers c and d, where d>1 and ceqpm1, every iterate in the zero orbit of f(z)=zd+c contains a primitive prime whenever zero has an infinite orbit. If c=pm1, then every iterate after the first contains a primitive prime.


The authors study primitive prime divisors in the orbit of zero under iterated application of a unicritical polynomial \(\varphi(z) = z^d+c\), with \(c,d \in \mathbb{Z}\). The main theorem states that, when \(d \geq 2\) and \(0\) is a wandering point of \(\varphi\), then \(\varphi^n(0)\) has a primitive prime divisor for all \(n \geq 1\), except of course for \(n = 1\) when \(c = \pm 1\). (Zero is not wandering exactly when \(c=0\), or \(c=-1\) and \(d\) is even, or \(c=-2\) and \(d=2\).) Here, a primitive prime divisor is a prime that divides \(\varphi^n(0)\) but not \(\varphi^k(0)\) for any \(k < n\). This improves a theorem of \textit{B. Rice} [Integers 7, No. 1, A26, 16 p. (2007; Zbl 1165.11028)]. The proof is based on Rice's observation that these sequences are rigid divisibility sequences.











This page was built for publication: Primitive prime divisors in zero orbits of polynomials

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2898399)