Finiteness theorems for vanishing cycles of formal schemes (Q891569)

Let \(k\) be a complete nonarchimedean valued field with a nontrivial valuation and let \(k^\circ\) be its ring of integers. The paper under review contains finiteness results for the étale cohomology of compact \(k\)-analytic spaces (in the sense of Berkovich) and for vanishing cycles on formal schemes over \(k^\circ\). We now give precise statements. Let \(X\) be a compact \(k\)-analytic space and let \(F\) be an abelian constructible sheaf on \(X\) with torsion orders prime to the residue characteristic. Assume that \(k\) is algebraically closed. Then, the groups \(H^q(X,F)\) are finite for \(q\geq 0\). Let \(\mathfrak{X}\) be a formal scheme locally topologically of finite presentation over \(k^\circ\) and let \(F\) be an abelian constructible sheaf on \(\mathfrak{X}_{\eta}\) with torsion orders prime to the residue characteristic. Then the vanishing cycles sheaves \(R^q \Psi_{\eta}(F)\) are constructible for \(q\geq 0\). The last results concern a more general class of formal schemes, namely special formal schemes, i.e. locally of the form Spf\((A)\), where \(A\) is a quotient of an algebra of the form \(k^\circ\{T_{1},\dots,T_{n}\}[[S_{1},\dots,S_{m}]]\). In this setting, the above result still holds if \(k\) is assumed to be discretely valued and if the constructibility condition is replaced by the more restrictive notion of \(\mathfrak{X}\)-constructibility. Let us mention that similar results were proven by \textit{R. Huber} [J. Algebr. Geom. 7, No. 2, 313--357 (1998; Zbl 1040.14008); J. Algebr. Geom. 7, No. 2, 359--403 (1998; Zbl 1013.14007)] under the assumption that \(k\) has characteristic 0. Note that \textit{V. Berkovich} himself had already proven similar results under algebraicity assumptions in [Invent. Math. 115, No. 3, 539--571 (1994; Zbl 0791.14008); ibid. 125, No. 2, 367--390 (1996; Zbl 0852.14002)]. Actually the proofs in the present paper rely on those former results, the extra hypotheses being removed thanks to Gabber's weak uniformization (see [\textit{L. Illusie} (ed.) et al., Travaux de Gabber sur l'uniformisation locale et la cohomologie étale des schémas quasi-excellents. Séminaire à l'École Polytechnique 2006--2008. Paris: Société Mathématique de France (SMF) (2014; Zbl 1297.14003)]) and Deligne's cohomological descent. Finally, as regards weak uniformization, let us mention that the author proves results of this kind in the settings he considers: \(k\)-analytic spaces (covered by generic fibers \(\hat{\mathcal{Y}}_{\eta}\) of schemes \(\mathcal{Y}\) over \(k^\circ\) with \(\mathcal{Y}_{\eta}\) smooth over \(k\)) and special formal schemes over \(k^\circ\) (covered by completions of semi-stable schemes over rings of integers of finite extensions of \(k\)). For this purpose, applying Gabber's results directly is sometimes not enough and the author needs to go back to the proof and adapt the arguments.