On \(L\)-functions of modular elliptic curves and certain \(K3\) surfaces (Q2115270)

A famous conjecture of \textit{D. H. Lehmer} [Duke Math. J. 14, 429--433 (1947; Zbl 0029.34502)] asserts that the Ramanujan tau function is never zero. The values of this function arise as the Fourier coefficients of the weight 12 modular form \[ \Delta(z)=q \prod_{n=1}^{\infty} \left( 1-q^{n} \right)^{24}=\sum_{n=1}^{\infty} \tau(n) q^{n}, \] where \(q=e^{2\pi i z}\). Recently, \textit{J. S. Balakrishnan} et al. [J. Number Theory 237, 3--14 (2022; Zbl 1497.11080)] proved that some odd integers of small absolute value cannot occur as a value of \(\tau(n)\). In forthcoming work, they were also able to obtain results for the Fourier coefficients of newforms of even weights \(\geq 4\) too. In this paper, the authors apply a similar method for weights \(k \leq 3\) and \(L\)-functions arising from weight \(2\) newforms and weight \(3\) newforms corresponding to modular elliptic curves. They are able to exclude some small odd values of the Fourier coefficients unconditionally, as well as being able to exclude more values under GRH. With the exceptions of \(\pm 81\) and \(\pm 93\), they also provide examples demonstrating that the values of the Fourier coefficients they are unable to eliminate under GRH do actually occur. In fact, it turns out that their use of GRH is unnecessary. All the results they prove under GRH can be shown to hold unconditionally. This comes from examining the Thue equations in their Table~4. If \(d \leq 37\), Pari/GP can unconditionally solve the Thue equations (albeit after approximately \(106,000\) seconds on a fast laptop for \(d=37\). For smaller \(d\), the time was far less.) For larger \(d\) and \(D=\pm d\), the method in the proof of Lemma~4.1 of their reference 5 applies. That only leaves \((d,D)=(41, \pm 83)\). Fortunately, neither of these two Thue equations has a solution modulo \(5\) or \(11\).