Hyperelliptic curves and newform coefficients (Q2030224)

Consider \(f\) a normalized newform of even weigh \(2k\) and level \(N\) with \(q\)-expansion \(f(z)=q+\sum_{n\geq 2} a_f(n)q^n\), such that \(a_f(n)\) are integers. The paper under review study which integers \(\alpha\) could appear at \(a_f(m)\) for some natural \(m\geq 2\). Lehmer conjectured, still unproven, that the weight 12 modular form \(\Delta(z)=\sum_{n\geq 1}\tau(n)q^n\) always satisfy \(\tau(n)\neq0\). Recent work has focused on the variant problem of showing that \(\tau(n)\) never takes certain values, see for example the works of \textit{M. R. Murty} et al. [Bull. Soc. Math. Fr. 115, 391--395 (1987; Zbl 0635.10020)], \textit{J. S. Balakrishnan} et al. in [``Variants of Lehmer's conjecture for Ramanujan's tau-function'', J. Number Theory (to appear)] and \textit{J. S. Balakrishnan} et al. [``Variants of Lehmer's speculation for newforms'', Preprint, \url{arXiv:2005.10354}]. From the work of [loc. cit.], \(a_f(n)\neq\pm\ell^m\) for a fixed odd prime \(\ell\), it is enough to study if \(a_f(p^{d-1})\neq \pm\ell^{m}\), where \(d\) is one of odd primes dividing \(\ell(\ell^2-1)\). For \(d\geq 7\), solutions to \(a_f(p^{d-1})=\pm\ell^m\) give solutions of Thue equations, which are well-studied. For \(d=3\), from the work in [loc. cit.], \(a_f(p^2)=\pm\ell^m\) are related with solutions of the hyperelliptic curves \(Y^2=X^{2k-1}+\pm\ell^m\) (already known when \(\ell^n\leq 100\) [\textit{Y. Bugeaud} et al., Compos. Math. 142, No. 1, 31--62 (2006; Zbl 1128.11013); \textit{C. F. Barros}, On the Lebesgue-Nagel equation and related subjects. University of Warwick (Ph.D. thesis) (2010)]), and for \(d=5\) are related with solutions of \(Y^2=5 X^{2(2k-1)}+4(\pm\ell^m)\). Again, inspired from [\textit{J. S. Balakrishnan} et al., ``Variants of Lehmer's speculation for newforms'', Preprint, \url{arXiv:2005.10354}], these equations for \(d=5\) are studied by Fibonacci type sequences as linear combination of the Fibonacci numbers with the Lucas numbers. In particular, in this paper the authors prove that for all odd primes \(\ell<100\) and all naturals we have \(\tau(n)\neq\pm\ell,\pm 5^m\). For general newforms \(f\) of even weight and level \(N\), the paper under review, obtained results of the flavour of Theorem 1.4 in [\textit{J. S. Balakrishnan} et al., ``Variants of Lehmer's speculation for newforms'', Preprint, \url{arXiv:2005.10354}], by use of modular method and a theorem of \textit{K. A. Ribet} on level-lowering [Int. Mat. Res. Not. 1991, No. 2, 15--19 (1991; Zbl 0728.11029)]. In particular if \(p\nmid a_f(p)\), then \(a_f(p^2)\neq m^j\) for any \(j\geq 4\) dividing \(2k-1\) and any nonzero integer \(m\), in particular, if \(2k\geq 6\), then \(a_f(p^2)\neq m^{2k-1}\). The authors obtains also a statement on \(a_f(p^4)\) under Frey-Mazur conjecture.