A refinement of Stark's conjecture over complex cubic number fields (Q1011661)

The authors study the first-order zero case of Stark's conjecture over a complex cubic number field \(F\). In that case, the conjecture predicts the absolute value of a complex unit in an abelian extension of \(F\). The authors present a refinement of Stark's conjecture by proposing a formula (up to root of unity), for the unit itself instead of its absolute value. Let \(E/F\) be a finite Galois extension of \(F\) with abelian Galois group \(G=\text{Gal}(E/F)\). According to class-field theory, \(G\) is canonically isomorphic (via the Artin reciprocity law) to a quotient of the narrow ray class group \(\text{Cl}_F^+(\mathfrak f)\) by one of its subgroups, where the conductor \(\mathfrak f\) of that group is a (finite) product of finite and infinite primes in \(F\) completely determined by the extension \(E/F\). It is assumed that \(G\) is isomorphic to the narrow ray class group \(\text{Cl}_F^+(\mathfrak f)\) and \(E=E_{\mathfrak f}\) is the corresponding ray class field. If \(\sigma\in G\), we write \(C\in \text{Cl}_F^+(\mathfrak f)\) for the corresponding ray class. Let the Dedekind zeta function \(\zeta_F(\sigma, s)= \zeta_F(C,s)= \sum_{\mathfrak a\in C} N (\mathfrak a)^{-s}\), \(\text{Re}(s)>1,\) where the sum runs over all the integral ideals \(\mathfrak a\) in the ideal class \(C\). Denote by \(F^{(1)} \subset \mathbb R\) the unique real embedding of \(F\) in \(\mathbb R\), and let \(F^{(2)}\in\mathbb C\) be one of the complex embeddings of \(F\) in \(\mathbb C\). Let \(E\hookrightarrow E^{(2)}\subset \mathbb C\) be an embedding of \(E\) into the field of complex numbers that extends the fixed complex embedding \(F\hookrightarrow F^{(2)}\subset \mathbb C\). Let \(w_E\) be the number of roots of unity in \(E\). Then Stark's conjecture [see \textit{D. S. Dummit, B. A. Tangedal} and \textit{P. B. van Wamelen}, Math. Comput. 73, No. 247, 1525--1546 (2004; Zbl 1046.11082)] is: There exists a unit \(\eta\in E^{(2)}\) \[ \zeta_F^\prime(\sigma,0)=\frac{d}{ds}\zeta_F(\sigma,s)|_{s=0} =- \frac{1}{w_E}\log \;|\sigma(\eta)|^2, \text{\;for\;all\;}\sigma\in G. \] Let \(U_{\mathfrak f}^+ = \{\varepsilon\in\mathbb Z_F^*\mid \varepsilon\equiv 1\pmod{\mathfrak f},\, \varepsilon^{(1)}>0\}\). The authors define successively the Shintani zeta function associated to a cone [\textit{T. Shintani}, J. Fac. Sci., Univ. Tokyo, Sect. I A 23, 393--417 (1976; Zbl 0349.12007)], a Shintani fundamental domain \(\mathfrak D\) for \(U_{\mathfrak f}^+\) consisting of a finite collection of open simplicial cones \(V_p \) (\(p\in S\) a finite index set), a corresponding set of associated rotations \(\mathfrak Z=\{T_p\mid p\in S\}\). By a succession of definitions and intermediate theorems, the authors define explicitly a complex number function \(\Theta(\mathfrak b, \mathfrak D,\mathfrak Z)\) depending only on an integral ideal \(\mathfrak b\) in the ray class \(C\), a fundamental Shintani domain \(\mathfrak D\) corresponding to \(\mathfrak b\), a corresponding set of rotations \(\mathfrak Z\) and an integer \(N(\mathfrak b,\mathfrak D)\) depending only on \(\mathfrak b\) and \(\mathfrak D\). The authors conjecture that: Let \(E=E_{\mathfrak f}\) be the narrow ray class field modulo \(\mathfrak f\), and let \(w_E\) be the number of roots of unity in \(E\). Then \[ \Theta(\mathfrak b, \mathfrak D, \mathfrak Z)^{w_E}=\xi(\mathfrak b, \mathfrak D,\mathfrak Z)\cdot\eta(C), \] where \(\eta(C)=\sigma(\eta)\) is a Stark's unit and \(\xi\) is a root of unity such that \(\xi^{N(\mathfrak b,\mathfrak D)}=1\), and they prove that: Up to multiplication by a root of unity, the complex number \(\theta(\mathfrak b,\mathfrak D,\mathfrak Z)\) depends only on the narrow ray class \(C\) containing the ideal \(\mathfrak b\). Observe that the conjecture is on Stark's units and not on their absolute values as Stark's conjecture. The conjecture is supported with some intermediate theorems and numerical examples.