Integral models of Hilbert modular varieties in the ramified case, deformations of modular Galois representations, and weight one forms (Q1717654)

The main result of this paper is the proof of strong Artin conjecture for odd continuous two-dimensional Galois representations \(\rho:\mathrm{Gal}(\bar{F}/F)\rightarrow \mathrm{GL}_2(\mathbb{C})\) of the absolute Galois group of a totally real number field \(F\); this is obtained by proving the icosahedral case of this conjecture, which was missing from previous works. For this, a number of strong results are proved on the geometry of Hilbert-Blumenthal varieties when \(p\) is ramified in \(F\). The strategy for proving the strong Artin Conjecture is similar to that of \textit{K. Buzzard} and \textit{R. Taylor} [Ann. Math. (2) 149, No. 3, 905--919 (1999; Zbl 0965.11019)] for \(F=\mathbb{Q}\): use Hida theory to pass information from weight \(2\) forms to weight \(1\) forms. To extend these ideas to the totally real case, the author is lead to define and study integral models of Hilbert-Blumenthal modular varieties at \(p\) when \(p\) is a prime which ramifies in the totally real number field \(F\). The author considers \([F:\mathbb{Q}]\)-dimensional moduli space of abelian varieties with real multiplication with PEL conditions and the cotangent bundle of the universal such object, first introduced in [\textit{M. Rapoport}, Compos. Math. 36, 255--335 (1978; Zbl 0386.14006)]. However, when \(p\) is ramified in \(F\), one has to work with integral models studied in [\textit{P. Deligne} and \textit{G. Pappas}, Compos. Math. 90, No. 1, 59--79 (1994; Zbl 0826.14027)]. Since abelian varieties \(A\rightarrow S\) in this case do not satisfy the condition that the Lie algebra of \(A\) is a free \(\mathcal{O}_F\otimes_{\mathbb{Z}}\mathcal{O}_S\)-module of rank \(1\) (this is called the \textit{Rapoport condition}), the idea of the author is to resolve the singularities of Deligne-Pappas models using ideas from [\textit{G. Pappas} and \textit{M. Rapoport}, J. Algebr. Geom. 12, No. 1, 107--145 (2003; Zbl 1063.14029)]. The main results of the paper can then be summarized as follows. Let \(G\) be the pull-back of \(\mathrm{Res}_{F/\mathbb{Q}}\mathrm{GL}_2\rightarrow\mathrm{Res}_{F/\mathbb{Q}}(\mathbb{G})\) along \(\mathbb{G}\rightarrow\mathrm{Res}_{F/\mathbb{Q}}(\mathbb{G})\), where \(\mathbb{G}\) is the multiplicative group scheme over \(F\). (1) Construction of integral models of \(G\) for level structure \(U\) which are maximal or Iwahori at primes of \(F\) dividing \(p\); these models admit descriptions as solution of moduli problems, which allow the author to introduce Katz \(p\)-adic modular forms in this context. These models have many application in the \(p\)-adic theory of modular forms (the work of \textit{C. Johansson} [Math. Ann. 357, No. 1, 51--88 (2013; Zbl 1346.11033)] on Coleman classicity theorem, possible extension to the ramified setting of the work of Andreatta-Iovita-Pilloni on the Hilbert eigenvariety, applications to Langlands functoriality in \(p\)-adic families). (2) Lifting theorems, showing that if \(\rho:\mathrm{Gal}(\bar{F}/F)\rightarrow\mathrm{GL}_2(\mathcal{O})\) a continuos, odd \(p\)-adic representation whose associated residual representation \(\bar\rho: \mathrm{Gal}(\bar{F}/F)\rightarrow \mathrm{GL}_2(k)\) is modular and satisfies suitable technical assumptions (among which the most relevant is probably the abslute irreducibility and the ordinary assumption at primes over \(p\)), then there exists a modular form \(f\) of weight \(1\) whose residual Galois representation coincides with \(\bar\rho\); this completes in the ramified setting previous works by several people, including the author, \textit{P. L. Kassaei} [J. Am. Math. Soc. 26, No. 1, 199--225 (2013; Zbl 1296.11052)], \textit{V. Pilloni} and \textit{B. Stroh} [Astérisque 382, 195--266 (2016; Zbl 1408.11052)]. The modular form \(f\) is a section of the universal automorphic bundle over the generic fiber at \(p\) of the compactification of the integral model of the Shimura variety constructed by the author. To prove this result, the author shows \(R=T\) theorems for Hida nearly ordinary deformations, and glue \(p\)-adic companion forms arising from morphisms \(T\rightarrow\mathcal{O}\) using a method of Kassaei [loc. cit.] to obtain a weight one classical form. One of the main technical problems solved by the author in this context is that, for \(A\) an abelian variety with real multiplication and PEL struction, and \(\mathfrak{p}\mid p\) a prime with \(p\) ramified in \(F\), the \(\mathfrak{p}\)-torsion \(A[\mathfrak{p}]\) of \(A\) is not a truncated Barsotti-Tate group (because the Rapoport condition fails in this case) and therefore the author needs to calculate the Dieudonné module with a new strategy. (3) Proof of the strong Artin conjecture in the remaining cases, thus completing the proof of this conjecture for totally real fields. The paper is extremely interesting and solves many difficult technical problems with new ideas.