Representing multiples of \(m\) in real quadratic fields as sums of squares (Q2112773): Difference between revisions

Let \(K\) be a totally positive number field, i.e., all infinite places are real, let \(\mathcal{O}\) be its ring of algebraic integers and denote by \(\mathcal{O}^+\) the subset of all totally positive integers. Let \(S(\mathcal{O})\) be the set of all sums of squares in \(\mathcal{O}\). Clearly, \(S(\mathcal{O})\subseteq \mathcal{O}^+\). A famous theorem by \textit{C. L. Siegel} [Math. Z. 11, 246--275 (1921; JFM 48.0179.04)] states that \(\mathbb{Q}\) and \(\mathbb{Q}(\sqrt{5})\) are the only totally positive number fields such that equality holds. In the present paper, the following variation of this problem is studied. Consider a real quadratic number field \(\mathbb{Q}(\sqrt{D})\) where \(D\geq 2\) is a square free integer, let \(m\in\mathbb{Z}\) with \(m>0\). For which \(D\) does one have that \(m\mathcal{O}^+ \subseteq S(\mathcal{O})\)? Siegel's theorem gives the answer for \(m=1\). Let \(\kappa=1\) (resp. \(=2\)) if \(D\equiv 1\bmod 4\) (resp. \(\equiv 2,3\bmod 4\)). \textit{V. Kala} and \textit{P. Yatsyna} [New York J. Math. 26, 1145--1154 (2020; Zbl 1460.11043)] have shown that for \(m=2\), inclusion holds iff \(D\in\{ 2,3,5\}\), and that \(m\mathcal{O}^+ \not\subseteq S(\mathcal{O})\) whenever \(m<\kappa\sqrt{D}/4\) or if \(m\) is odd and \(D\equiv 2,3\bmod 4\), and that if \(m\geq D/2\) then each element in \(\kappa m\mathcal{O}^+\) is a sum of five squares in \(\mathcal{O}\). The author obtains a refinement of the above bounds by providing various explicit intervals that assure that inclusion does not hold whenever \(m\) falls into one of these intervals. The author furthermore shows that for \(m=4\), inclusion holds if and only if \(D\in\{ 2, 3, 5, 6, 7, 10, 11, 13\}\). For certain classes of integers \(D\), the author also develops an algorithm that allows to determine all \(m\) for which inclusion holds.