Lucas and Lehmer numbers without primitive divisor (Q873831): Difference between revisions

Let \(\alpha\) and \(\beta\) be such that \(\alpha+\beta\) and \(\alpha\beta\) are coprime nonzero integers. Assume further that \(\alpha/\beta\) is not a root of \(1\). The numbers \(u_n=(\alpha^n-\beta^n)/(\alpha-\beta)\) are called Lucas numbers. A variant of Lucas numbers was introduced by Lehmer and since then such numbers are called Lehmer numbers. A primitive divisor of a Lucas or Lehmer number \(u_n\) is a prime factor \(p\) of \(u_n\) which does not divide neither the discriminant \((\alpha-\beta)^2\) of the quadratic equation having \(\alpha\) and \(\beta\) as roots, nor \(u_m\) for any \(1\leq m<n\). Primitive divisors are important since they are large (they are at least as large as \(n-1\)). Furthermore, they have found applications in the effective resolution of certain types of exponential Diophantine equations. The existence of primitive divisors for Lucas and Lehmer numbers initiated in the late 19th century in a paper of \textit{K. Zsigmondy} [ On the theory of power residues (German). Monatsh. Math. 3, 265--284 (1892; JFM 24.0176.02)]. After several contributions by Carmichael, Lehmer, Stewart and Schinzel, it culminated with the work of \textit{Y. Bilu, G. Hanrot} and \textit{P. M. Voutier} [J. Reine Angew. Math. 539, 75--122 (2001; Zbl 0995.11010)]. The final theorem is that if \(n\not\in \{1,2,3,4,5,6,8,10,12\}\), then the \(n\)th term of a Lucas or Lehmer sequence has a primitive divisor except for finitely many triples \((n,\alpha,\beta)\), all of which are known. However, for the small values of \(n\) mentioned above there are parametric families of triples \((n,\alpha,\beta)\) for which the \(n\)th Lucas or Lehmer number has no primitive divisors. This list was first compiled by \textit{P. M. Voutier} [ Math. Comput. 64, 869--888 (1995; Zbl 0832.11009)], although a few cases were missing. In this paper, the author reworks the arguments of the primitive divisor theorem for the above small values of \(n\) and completes Voutier's list.