Artin's conjecture for Drinfeld modules (Q2166198)

Artin's conjecture states that if \(a\in{\mathbb Z}\), \(a\neq0,\pm 1\) or any perfect square, then the density of the primes such that \(a\bmod p\) generates \(({\mathbb Z}/p{\mathbb Z})^*\) is positive. Let \(E\) be an elliptic curve defined over \({\mathbb Q}\) and \(a\in E({\mathbb Q})\) be a point of infinite order. Then we say that \(a\) is a primitive point for \(E\) modulo \(p\) if \(E\) has good reduction at \(p\) and the residue of \(a\) in \(E\) modulo \(p\) generates the finite group \(E\) modulo \(p\). \textit{R. Gupta} and \textit{M. R. Murty} [Compos. Math. 58, 13--44 (1986; Zbl 0598.14018)] proved, assuming the generalized Riemann hypothesis, that if \(E\) is an elliptic curve over \({\mathbb Q}\) with complex multiplication by the ring of integers in a quadratic imaginary extension \(L\) of \({\mathbb Q}\), \(a\) is a point of infinite order and \(M_a(x)\) is the number of primes \(p\leq x\) such that \(p\) splits in \(L\) and \(a\) is a primitive point modulo \(p\), then \[ M_a(x)=C_E(a)\mathrm{li}(x)+O\big(\frac{x\log\log x}{(\log x)^2}\big) \] as \(x\to\infty\) and \(C_E(a)>0\) if \(2\) and \(3\) are inert in \(L\) or \(L= {\mathbb Q}(\sqrt{-11})\). \textit{C.-N. Hsu} and \textit{J. Yu} [J. Number Theory 88, No. 1, 157--174 (2001; Zbl 1090.11039)] proved the following. Let \(k\) be a global function field over \({\mathbb F}_r\) and let \(A\) be the ring of elements of \(k\) regular away form a place \(\infty\) of degree \(1\). Let \(H\) be the Hilbert class field of \(A\), and let \({\mathcal O}\) be the integral closure of \(A\) in \(H\). Let \(\mathrm{sgn}\) be a sign function, \(\psi:A\to {\mathcal O}\{\tau\}\) be a \(\mathrm{sgn}\)-normalized Drinfeld module of rank \(1\). For \(0\neq a\in{\mathcal O}\) and \(r\neq 2\), let \(N_a(x)\) be the number of primes \({\mathfrak p} \in {\mathcal O}\) of degree \(x\) such that \(a+{\mathfrak p}\) generates \({\mathcal O}/{\mathfrak p}\) as an \(A\)-module. Then \[ N_a(x)=\delta_a \frac{r^x}{x}+o\big(\frac{r^x}{x}\big) \] and \(\delta_a\neq 0\) is given by an Euler product. The arguments in both, Gupta and Murty's paper and Hsu and Yu's paper do not work for nonprincipal ideal domains. In the present paper, the authors prove a Drinfeld module analogue to Gupta and Murty's result, or a rank \(2\) analogue of Hsu and Yu's result. As a corollary, they conclude Artin's conjecture for rank \(1\) Drinfeld modules \(\psi:B\to K\{\tau\}\) where \(K\) is any finite extension of \(k\), and \(a\) is a nontorsion element of \(K\) for which the division fields \(K({\mathfrak q}^{-1}\langle a\rangle)\) are not equal to \(K\) for every prime ideal \({\mathfrak q}\) of \(B\). The main result is the following. Let \(F\) be a global function field over \({\mathbb F}_r\), \(\infty\) a place of \(F\), \(A\) the ring of elements of \(F\) regular away form \(\infty\), and \(K\) a finite extension of \(F\). Let \(\phi:A\to K\{\tau\}\) be a Drinfeld module of generic characteristic and of rank \(2\). Let \(B=\mathrm{End}(\phi)\) and assume that \(B\) is the integral closure of \(A\) in a quadratic extension of \(F\) and that the elements of \(B\) have coefficients in \(K\). For \(a\in K\), let \(N_a(x)\) be the number of the primes \(\wp\) of \(K\) of degree \(x\) and such that the reduction of \(a\) generates \({\mathbb F}_{\wp}\) as an \(A\)-module. Then there exist a positive integer \(J\) and constants \(\delta(0),\ldots,\delta(J-1)\), depending on \(\phi\) and \(a\) such that as \(x\equiv i\bmod J\) and \(x\to\infty\), we have \[ N_a(x)=\delta(i)\frac {r^x}{x}+O\big(\frac{r^x\log x}{x^2}\big). \]