On rotated \(D_n\)-lattices constructed via totally real number fields (Q1949399)

A lattice \(\Lambda\) is a discrete additive subgroup of \(\mathbb R^n\) considered as the standard real vector space. Equivalently, \(\Lambda\subset \mathbb R^n\) is a lattice if and only if there are linearly independent vectors \(v_1,\dots,v_n\in\mathbb R^n\) such that \(\Lambda = \sum_{i=1}^m a_iv_i, a_i\in\mathbb Z\). For \(n \geq 3\), the \(n\)-dimensional lattice \(D_n\) in \(\mathbb R^n\) is \[ D_n = \{(x_1,\dots, x_n) \in \mathbb Z^n; \sum_{i=1}^n x_i\text{\;is even\;}\}. \] A lattice \(\Lambda\) has diversity \(k \leq n\) if \(k\) is the maximum number such that for all \(0\not = \mathbf y \in \Lambda\), there are at least \(k\) nonzero coordinates, and \textit{full diversity} if \(k=n\). Full diversity \textit{rotated \(D_n\)-lattices} are obtained in [\textit{G. C. Jorge} et al., J. Number Theory 132, No. 11, 2397--2406 (2012; Zbl 1272.11084)] using subfields of some cyclotomic fields, see also \textit{E. Bayer-Fluckiger} et al. [IEEE Trans. Inf. Theory 50, No. 4, 702--714 (2004; Zbl 1182.94034)]. Let \(K\) be a totally real number field of degree \(n = [K : \mathbb Q]\) and \(\mathcal O_K\) its ring of integers. Let \(\{\sigma_i\}_{i=1}^n\) be the real embeddings. The embedding \(\sigma_\alpha : K \rightarrow \mathbb R^n\) where \(\sigma_\alpha(x) =(\sqrt{\alpha_1}\sigma_1(x),\dots,\sqrt{\alpha_n}\sigma_n(x))\) is called the \textit{twisted embedding}. Let \(\zeta_n\) denote the \(n\)th primitive root of unity. In this paper, considering the compositum \(K_3\) of the fields \(\mathbb Q(\zeta_{2^r}+\zeta_{2^r}^{-1})\) and \(\mathbb Q(\zeta_{p_1}+\zeta_{p_1}^{-1})\) and the compositum \(K_4\) of \(\mathbb Q(\zeta_{p_1} + \zeta_{p_1}^{-1})\) and \(\mathbb Q(\zeta_{p_2} + \zeta_{p_2}^{-1})\), where \(p_1\) and \(p_2\) are prime numbers with \(p_1 \not= p_2\), the authors construct some new families of rotated \(D_n\)-lattices via free \(\mathbb Z\)-modules of rank \(n\) that are not ideals. Then they show that it is impossible to construct these lattices via fractional ideals of \(\mathcal O_{K_i}\) the ring of integers of \(K_i\), for \(i = 2, 3, 4\). Lastly, they present a necessary condition for constructing a rotated \(D_n\)-lattice via a totally real Galois extension \(K/\mathbb Q\). These results are related to an open problem stated in [\textit{E. Bayer-Fluckiger}, in: Algebraic geometry: Hirzebruch 70, Contemp. Math. 241, 69--84 (1999; Zbl 0951.11016)] p. 69: given a lattice \(\Lambda\), is there a number field K such that it is possible to construct via \(\mathcal O_K\)? More precisely, we give an explicit example of their approach. Let us denote \(e_0 = 1, e_i = \zeta_{2^r}^i+\zeta_{2^r}^{-i}\) for \(i=1,\dots,2^{r-2}-1\) and \(b_i=\zeta_p^i+\zeta_p^{-i}\) for \(i=1,\dots,\frac{p-1}{2}\). The authors prove: Consider \(K = K_1K_2\) the compositum of \(K_1\) and \(K_2\), where \(K_1=\mathbb Q(\zeta_{2^r}+\zeta_{2^r}^{-1})\) and \(K_2=\mathbb Q(\zeta_p+\zeta_p^{-1})\) for \(r \geq 3\) and \(p \geq 5\) prime. Let \(n_1 = 2^{r-2}\) and \(n_2 = \frac{p-1}{2}\) . If \(\mathcal I\) is the \(\mathbb Z\)-submodule of \(\mathcal O_K\) with \(\mathbb Z\)-basis \[ \gamma = \{e_0b_1,\dots, e_0b_{n_2-1}, 2e_0b_{n_2}, e_1b_1,\dots, e_1b_{n_2},\dots, e_{n_1-1}b_1,\dots, e_{n_1-1}b_{n_2}\}, \] then the lattice \((\sqrt{2^{r-1}p})^{-1}\sigma_\alpha(\mathcal I)\subseteq \mathbb R^n\), where \(\alpha = (2-e_1)(2-b_1)\), is a rotated \(D_n\)-lattice.