A prime-to-\(p\) version of Grothendieck's anabelian conjecture for hyperbolic curves over finite fields of characteristic \(p>0\) (Q1016585)

Grothendieck's anabelian conjecture for smooth, hyperbolic curves over finite fields states that, given any two hyperbolic curves \(U\) and \(V\) over finite fields, any isomorphism between the arithmetic fundamental groups \(\pi_1(U)\) and \(\pi_1(V)\) arises from a unique isomorphism \(U \to V\). In particular, if \(U\) is a smooth hyperbolic curve over a finite field, the fundamental group of \(U\) determines the isomorphism class of \(U\). This was proven in \textit{A. Tamagawa} [Compositio Math. 109, 135--194 (1997; Zbl 0899.14007)] in the affine case and in \textit{S. Mochizuki} [J. Math. Kyoto Univ. 47 (2007; Zbl 1143.14305)] in the proper case. The current paper proves a stronger, ``prime-to-\(p\)'' version of this result: For any smooth curve \(X\) over a finite field \(k_X\), there is a natural inclusion \(\pi_1(\overline{X} := X \times_{\text{Spec } k_X} \text{Spec } \overline{k_X}) \hookrightarrow \pi_1(X)\). Let \(\Delta_X\) be the maximal prime-to-\(p\) quotient of \(\pi_1(\overline{X})\), and let \(S\) be the kernel of the canonical map \(\pi_1(\overline{X}) \to \Delta_X\). Write \(\Pi_X := \pi_1(X)/S\). Then, if \(U\) and \(V\) are smooth hyperbolic curves defined over finite fields, any isomorphism \(\Pi_U \to \Pi_V\) arises from a unique isomorphism \(U \to V\). In particular, \(\Pi_U\) determines the isomorphism class of \(U\). A similar birational result is also proved for absolute Galois groups of function fields, strengthening a result of \textit{K. Uchida} [Ann. Math. 106 (1977; Zbl 0372.12017)]. The outline of the proof is similar to that of \textit{A. Tamagawa} [Compositio Math. 109, 135--194 (1997; Zbl 0899.14007)]. Given \(\Pi_X\), one first addresses the ``local theory'', attempting to recover the closed points of \(X\) by recovering the conjugacy classes of decomposition groups of closed points of \(X\) as subgroups of \(\Pi_X\). In the case treated by the current paper, the points of \(X\) can only be recovered up to a finite set \(E_X\), the ``exceptional set''. Then, one uses Kummer theory to recover the group \(\mathcal{O}_{E_X}^{\times}\) of rational functions on \(X\) with support away from \(E_X\). Lastly, one recovers the additive structure on the set of rational functions \(k_X(X)\). The problem of the exceptional set \(E_X\) is dealt with by a trick allowing one to pass to an infinite extension of \(k_X\) over which \(X\) has infinitely many points, followed by a descent argument.