Sporadic cubic torsion (Q2239246)

Aside from the main result, the paper could easily be taken as a survey of known results in the classification of finite groups that appear as the torsion subgroup of an elliptic curve \(E(K)\) where \(K\) is a cubic number field. The main result proves a conjecture that classifies these finite groups which was listed by [\textit{D. Jeon} et al., Acta Arith. 113, No. 3, 291--301 (2004; Zbl 1083.11038)] and is also officially stated in [\textit{J. Wang}, On the torsion structure of elliptic curves over cubic number fields. University of Southern California (PhD thesis) (2015)]: A finite abelian group is a torsion group of an elliptic curve over a cubic number field iff it is one of the following (where \(C_n\) is the cyclic group of order \(n\)): \[ \begin{array}{ll} C_n & n=1,\dots,16,18,20,21\\ C_2 \oplus C_m & m=1,\dots,7 \end{array} \] The upper bound for the prime divisor of the order of a torsion point in such elliptic curves was proven by [\textit{P. Parent}, J. Théor. Nombres Bordx. 15, No. 3, 831--838 (2003; Zbl 1072.11037)] (\(p\le 13\)), combining this and the past results by Wang, Jeon et al., Bruin and Najman the remaing task was to prove \[n=22,24,25,26,27,30,32,33,35,39,45,65,121\] and \(m=8,9\) cannot occur (the main objective of the paper) and to finally prove the conjecture. The result very much depends on direct computations (all proofs are based on computations). These are done mainly by Maple, Sage and most notably by Magma. The authors do not directly provide the source code for their computation, however there is a github repository for their codes: \url{https://github.com/jmorrow4692/SporadicCubicTorsion}. To prove the above, the authors had to determine the cubic points of the modular curve \(X_1(n)\). Different techniques are used for this, depending on whether the Jacobian of these curves have rank \(0\) or whether it has positive rank. The authors show (Theorem 3.1) that only for \(n=65,121\) (in the above list), we will have positive rank for \(J_1(n)(\mathbb Q)\). For modular Jacobians of rank \(0\) different techniques such as \emph{direct analysis over \(\mathbb Q\)}, \emph{direct analysis over \(\mathbb F_p\)}, etc. are used. For modular Jacobian of positive rank (\(n=65,121\)) the formal immersion criterion is used. Finally, the authors justify why they needed to investigate the cases \(n=22,25,65\), despite the fact that Wang addressed these cases in the past. They claim that there were gaps in the argument of Wang which they explained in detail in Remark 7.5 of the paper.