scientific article; zbMATH DE number 1033059

Let \(R\) be a Dedekind domain with quotient field \(K\). From the early 1960's, it is known that any abelian variety \({\mathcal A}_K\) defined over \(K\) admits a good model over \(R\), the so-called Néron model. This is a smooth (separated) \(R\)-scheme of finite type \({\mathcal A}_R\) which satisfies the condition \({\mathcal A}_K={\mathcal A}_R \times {_RK}\) and the following universal property: (N) Given a smooth \(R\)-scheme \({\mathcal Z}_R\) and a morphism of \(K\)-schemes \(u: {\mathcal Z}_K \to {\mathcal A}_K\), \(u\) extends uniquely to a morphism of \(R\)-schemes \({\mathcal Z}_R \to {\mathcal A}_R\). The notion of Néron model can be generalized to any smooth, separated \(K\)-scheme of finite type \({\mathcal X}_K \) but its existence is not in general assured. For many purposes one finds it useful to switch from the scheme setting to the formal/rigid one. Then one restricts to Dedekind domains \(R\) which are complete discrete valuation rings. One defines then a formal Néron model of a smooth rigid \(K\)-space \(X_K\). This is a smooth formal \(R\)-scheme \(U_R\), whose generic fibre \(U_{R,K}\) is an open rigid subspace of \(X_K\) and which satisfies the following universal property: \((N')\) Given a smooth formal \(R\)-scheme \(Z_R\) and a morphism of rigid \(K\)-spaces \(u_K: Z_{R,K} \to X_K\), \(u_K\) extends uniquely to a morphism of formal \(R\)-schemes \(Z_R \to U_R\). One should observe that the formal Néron model \(U_R\) is not required to be a true model of \(X_K\) because in the formal/rigid setting the relation between a formal scheme and its special fibre is so strong that we had really few possibilities to find such a model. Although one works with this weaker condition, the existence of formal Néron models is not in general assured. One has to restrict to quasicompact smooth rigid \(K\)-groups in order to find a family of rigid \(K\)-spaces for which the search is always successful. Formal Néron models arise quite naturally as completions of usual Néron models. By this we mean that taken a smooth \(K\)-group scheme of finite type \({\mathcal X}\) (resp. an abelian variety) and considered the associated rigid \(K\)-group \(X_K= {\mathcal X}^{an}\), then the formal completion of the Néron model of \({\mathcal X}\) along its special fibre is a Néron quasi-model (resp. a Néron model) of \(X_K\). A Néron quasi-model of \(X_K\) is a formal \(R\)-scheme \(V_R\) which satisfies \((N')\) and whose canonical morphism \(V_{R,K}\to X_K\) is an open immersion when restricted to the open quasi-compact subsets of \(V_{R,K}\). It is clear that also the formation of formal Néron models is compatible with étale base change of the valuation ring, i.e. with finite unramified extension of \(K\). We are interested to understand their behaviour under more general base change. To do this, we introduce ``Weil restriction functors also on the category of rigid \(K\)-spaces and on the category of formal \(R\)-schemes. Suppose then given a finite field extension \(K'\) of \(K\) and a rigid \(K\)-group \(X_K\) such that \(X_{K'} =X_{K \times_K} K'\) admits a formal Néron model \(U_{R'}'\) over \(R'\). We show that the formal Néron model \(U_R\) of \(X_K\) can be constructed as a group smoothening of the schematic closure of \(H_K= X_K \times_{{ \mathfrak R}_{K'/K} (X_{K'})} {\mathfrak R}_{R'/R} (U_{R'}')_K\) in the formal ``Weil restriction \({\mathfrak R}_{R'/R} (U_{R'}')\). Moreover, if we suppose the extension to be tamely ramified, we can extend results of \textit{B. Edixhoven} [Compos. Math. 81, No. 3, 291-306 (1992; Zbl 0759.14033)], proving that the schematic closure defined above is already smooth and we can avoid to apply the smoothening process. The statement about formal Néron models sounds more complicated than the analogous in the scheme context because the generic fibre of a formal Néron model is only an open subspace of the whole space and the generic fibre of the formal ``Weil restriction scheme is smaller than the rigid ``Weil restriction space. For these reasons we have to devote a great part of this thesis to the study of the rigid ``Weil restriction and its relation with the formal one.

Recommendations

Cited in

(11)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4344089)