A family of pairs of imaginary cyclic fields of degree \((p-1)/2\) with both class numbers divisible by \(p\) (Q1987614)

From MaRDI portal
Revision as of 04:22, 30 July 2023 by Importer (talk | contribs) (‎Created a new Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
A family of pairs of imaginary cyclic fields of degree \((p-1)/2\) with both class numbers divisible by \(p\)
scientific article

    Statements

    A family of pairs of imaginary cyclic fields of degree \((p-1)/2\) with both class numbers divisible by \(p\) (English)
    0 references
    0 references
    0 references
    15 April 2020
    0 references
    In their previous paper [J. Number Theory 176, 333--343 (2017; Zbl 1422.11221)] the authors of the paper under review proved the following theorem where \(F_n\) is the \(n\)-th Fibonacci number. Theorem 1. For \(n\in {\mathcal N} := \{n\in {\mathbb N} \mid n\equiv \pm 3\pmod{500}, n\not\equiv 0 \pmod 3\}\), the class numbers of both \({\mathbb Q}(\sqrt{2-F_n})\) and \({\mathbb Q}(\sqrt{5(2-F_n)})\) are divisible by \(5\). Moreover, the set of pairs \[ \left \{({\mathbb Q}(\sqrt{2-F_n}), {\mathbb Q}(\sqrt{5(2-F_n)}))\Big |n\in {\mathcal N}\right\} \] is infinite. In the present paper they obtain a result of which Theorem 1 is a special case. To state this result let \(p\) be a prime number such that \(p\equiv 5 \pmod 8\) and let \(\zeta\) be a primitive \(p\)-th root of unity. Let \(\delta\) be a generator of \(\mathrm{Gal}({\mathbb Q} (\zeta)/{\mathbb Q})\) and put \(\delta_0 := \delta^{(p-1)/4}\). Let \(\omega_0 := \zeta + \zeta^{\delta_0} + \zeta^{\delta_0^2} + \zeta^{\delta_0^3}\). Then \(k_0 := {\mathbb Q}(\omega_0)\) is the unique subfield of \({\mathbb Q} (\zeta)\) of degree \((p-1)/4\). Let \(u_p > 1\) be the fundamental unit of \(k={\mathbb Q}(\sqrt{p})\). Then \[ u_p=\frac{t+b\sqrt{p}}{2} \] where \(t,b\in {\mathbb Z}\) with \(t, b >0\). Now, using the trace \(t\) of \(u_p\), define the following two sequences \(\{{\mathcal F}_n\}\) and \(\{{\mathcal L}_n\}\) by \[ \begin{cases} {\mathcal F}_0 := 0,\ {\mathcal F}_1 := 1, \ {\mathcal F}_{n+2} := t{\mathcal F}_{n+1} +{\mathcal F}_n \ (n\in {\mathbb Z}),\\ \\ {\mathcal L}_0 := 2,\ {\mathcal L}_1 := t, \ {\mathcal L}_{n+2} := t{\mathcal L}_{n+1} +{\mathcal L}_n \ (n\in {\mathbb Z}). \end{cases} \] For integers \(m\) and \(n\), and a prime number \(q\) \((\neq p)\), put \[ D_{m,n} := {\mathcal L}_m(2{\mathcal F}_m - {\mathcal F}_n{\mathcal L}_m)b, \] \[ N_q := \begin{cases} \mathrm{lcm}(p^2(p-1), q-1) \text{ if } \left (\frac{p}{q}\right )=1,\\ \mathrm{lcm}(p^2(p-1), 2(q+1))\text{ if } \left (\frac{p}{q}\right)=-1. \end{cases} \] (As the authors point out, when \(m\) and \(n\) are odd and \(n>3\), \(D_{m,n}\) is negative since \({\mathcal F}_{-m} = (-1)^{m+1}{\mathcal F}_m\) and \({\mathcal L}_{-m} = (-1)^m{\mathcal L}_m\)). Finally, put \[ \alpha = \alpha(m,n) := \frac{{\mathcal L}_n {\mathcal L}_m + ({\mathcal L}_m{\mathcal F}_n - 2{\mathcal F}_m)b\sqrt{p}}{2}, \] \[ f_{\alpha}(X) := X^4 -TX^3+(N+2)X^2 - TX + 1, \] \[ f_{\alpha, q}(X) := f_{\alpha} \bmod q \in {\mathbb F}_q[X], \] where \(N := N_{k/{\mathbb Q}}(\alpha)\) and \(T :=\mathrm{Tr}_{k/{\mathbb Q}}(\alpha)\). The first main result the authors obtain is Main Theorem 1. We assume that there exist integers \(m_0,n_0\) with \(m_0\equiv n_0 \equiv 1 \pmod 2\) and a prime number \(q\) such that \begin{itemize} \item[(i)] \(({\mathcal L}_{m_0}{\mathcal F}_{n_0} - 2{\mathcal F}_{m_0})b \equiv 0 \pmod {p^2}\), \item[(ii)] \(q \nmid 2bp\) and \(f_{\alpha_0,q}(a) = 0\) for some \(i\in \{1,2,4\}\) and \(a\in {\mathbb F}_{q^i}\setminus {\mathbb F}^p_{q^i}\), where \(\alpha_0 := \alpha(m_0,n_0)\). \end{itemize} Then for any pair \((m,n)\in {\mathcal N} := \left \{(m,n)\in {\mathbb Z}^2 \Big | m\equiv m_0 \pmod {N_q}, n\equiv n_0 \pmod{N_q}, n>3 \right \}\), the class numbers of both imaginary cyclic fields \(k_0(\sqrt{D_{m,n}})\) and \(k_0(\sqrt{pD_{m,n}})\) of degre \((p-1)/2\) are divisible by \(p\). Moreover, the set of pairs \[ \left \{(k_0(\sqrt{D_{m,n}}), k_0(\sqrt{pD_{m,n}})) \Big |(m,n)\in {\mathcal N}\right \} \] is infinite. The second main result of the paper is Main Theorem 2. Assume that ERH holds. Then there exist the integers \(m_0, n_0\) and the prime number \(q\) as in Main Theorem 1. Here ``ERH'' means the extended Riemann hypothesis for \(k(\zeta_n, \sqrt [n]{u_p})\) for every square free integer \(n>0\). The authors conclude their paper with some examples illustrating Main Theorem 1. One of these shows how Main Theorem 1 implies Theorem 1.
    0 references
    class numbers
    0 references
    abelian number fields
    0 references
    fundamental units
    0 references
    Gauss sums
    0 references
    Jacobi sums
    0 references
    linear recurrence sequences
    0 references

    Identifiers