Multiples of integral points on elliptic curves (Q2378051): Difference between revisions

Let \(E/\mathbb{Q}\) be a minimal elliptic curve and \(P=(x_P,y_P)\in E(\mathbb{Q})\) a point of infinite order. For any prime \(p\) let \(r(P,p)\) be the order of \(P\) in the group \(E(\mathbb{Q}_p)/E_0(\mathbb{Q}_p)\) (which is trivial for all primes of good reduction) and let \(M(P)=\text{lcm}\{r(P,p)\,:\,p\;\text{prime}\}\). The paper deals with integers \(n\) such that \(nP\) is an integral point and gives a bound for all such \(n\)'s (except possibly one) in terms of the height of \(E\) and \(M(P)\). The author provides explicit bounds for the denominators of the \(x_{nP}\)'s via division polynomials and uses a result of \textit{S. David} [Mém. Soc. Math. Fr., Nouv. Sér. 62 (1995; Zbl 0859.11048)], which bounds linear forms in elliptic logarithms, to prove that there is at most one value \(n > CM(P)^{16}\) such that \(nP\) is integral (\(C\) is an absolute constant). It is worth noticing that such unique ``large'' exception, if it exists, is prime. In the final section, combining the previous main theorem with results of the author and J. Silverman, the author examines quasi-minimal quadratic twists of elliptic curves and, in particular, the curves \(E_N\,:\,y^2=x^3-N^2x\) with a square-free \(N\). The bound for a non-torsion point \(P\in E_N(\mathbb{Q})\) can be reduced to 1, i.e., there is at most one natural number \(n>1\) such that \(nP\) is integral (the proof is obtained by examining ``large'' \(N\) at first and then dealing with the finitely many remaining curves by direct computation).