\(D(n)\)-quintuples with square elements (Q1982632): Difference between revisions

Let \((\mathbb{Q}^\times)^2\) denote the set of rational squares. An \(m\)-tuple \((x_1,\dots,x_m)\in \mathbb{Q}^m\) is said to have property~\(D(a)\) if the conditions \(x_ix_j(x_i-x_j)\neq 0\) and \(x_ix_j+a\in (\mathbb{Q}^\times)^2\) hold for all pairs \(\{i,j\}\in \binom{[m]}{2}\). When \(a=1\), such an \(m\)-tuple is also called a rational Diophantine \(m\)-tuple. Call a vector \((v_1,\dots,v_m)\in \mathbb{Z}^m\) primitive if \(\gcd(v_1,\dots,v_m) = 1\). Note that a primitive vector in \(\mathbb{Z}^m\), with not all coordinates negative, has property~\(D(0)\) if and only if each of its coordinates is a square. The authors show that there are infinitely many rational Diophantine quintuples with property~\(D(0)\); or equivalently, that there are infinitely many primitive vectors \(v\in \mathbb{Z}^5\) such that the following hold simultaneously: (i) \(v\) has property \(D(u^2)\) for some integer \(u\geq 1\), and (ii) each coordinate of \(v\) is a square. (See Theorem~1 and the ensuing paragraph.) The proof is the result of a clever mix of theory and experimentation. The initial motivation for the paper was the ``experimental discovery'' of a rational Diophantine quintuple \((a,b,c,d,e)\) with the following properties: (i) \(abcd=1\), (ii) \((a+b-d-e)^2 = 4(ab+1)(de+1)\) and \((a+c-d-e)^2 = 4(ac+1)(de+1)\), and (iii) \((a,b,c,d,e)\) has property~\(D(0)\). (See the last paragraph of \S1, and the first of \S2.) The authors call such quintuples ``exotic'' (see Definition~1). They relate exotic quintuples directly to the rational points on an explicit general-type surface \(S\) (see Proposition~2, Remark~1, and \S2), outside a certain ``degeneracy locus'' (see the first sentence of \S3). In view of the Bombieri-Lang philosophy, it is then remarkable that the authors -- using a clever computer search inspired by Riemann-Hurwitz, Lemma~5, and the ``large symmetry group'' of a certain branch locus under \(S\) -- are able to construct genus-one curves on \(S\) giving infinitely many exotic quintuples (see Proposition~3 and \S\S3--4). In the final \S5, the authors make several interesting remarks: ``it remains open if there is a rational Diophantine quintuple with square elements''; ``there are infinitely many rational Diophantine quadruples with square elements, for example the following two parameter family\dots''; and ``[there is also] an example of a rational Diophantine quadruple with square elements for which the product \(abcd\neq 1\) [given as follows]''.