Supersingular \(K3\) surfaces for large primes. With an Appendix by Andrew Snowden. (Q465458): Difference between revisions

Let \(k\) be an algebraically closed field of characteristic \(p>0\). Let \(X\) be a \(K3\) surface defined over \(k\). A \(K3\) surface \(X\) is said to be \textit{supersingular} if its formal Brauer group \(\widehat{\text{{Br}}}(X)\) has infinite height, i.e., \(\widehat{\text{{Br}}}(X)=\widehat{{\mathbb{G}}_a}\). Artin conjectured that a supersingular \(K3\) surface over \(k\) has the Picard rank \(22\). \textit{M. Artin} [Ann. Sci. Éc. Norm. Supér. (4) 7, 543--567 (1974; Zbl 0322.14014)] proved his conjecture when \(X\) admits an elliptic fibration. The main result of this article is to establish the Artin conjecture under certain conditions, and is formulated as follows. Theorem: Let \(k\) be an algebraically closed field of characteristic \(p>0\). If \(X\) is a supersingular \(K3\) surface over \(k\) admitting a polarization of degree \(2d\) with \(p>2d+4\), then the Artin conjecture holds for \(X\). Let \(p\geq 5\), and assume semistable reduction for smooth projective surfaces over discrete valuation fields with residue field \(k\), then the Artin conjecture holds for a supersingular \(K3\) surface \(X\) which admits a polarization of degree prime to \(p\). Here a semistable reduction over a discrete valuation field means that, after a finite base change, there is a semistable model over the valuation ring. Now assume that \(k=\bar{\mathbb{F}}_p\). A \(K3\) surface \(X\) is said to satisfy the Tate conjecture about algebraic cycles if, for every model \(X^{\prime}\) defined over \(\mathbb{F}_{p^r}\), the map \(\text{{Pic}}(X^{\prime})\otimes\mathbb{Q}_{\ell} \to H^2_{\text{ét}}(X,\mathbb{Q}_{\ell}(1))^{\text{{Gal}}(k/\mathbb{F}_{p^r})}\) is an isomorphism. As a corollary of the main result, one obtains the following: Corollary. The Tate conjecture holds for a \(K3\) surface admitting a polarization of degree \(2d\) such that \(p>2d+4\). Assuming semistable reduction and \(p\geq 5\), the Tate conjecture holds for \(K3\) surfaces admitting a polarization of degree prime to \(p\). The basic idea of proof is the following. In a connected family of supersingular \(K3\) surfaces, the rank of the Picard group must remain constant. Using this fact and the valid Artin's conjecture for elliptic \(K3\) surfaces, it then suffices to show that every connected component of the supersingular locus in the moduli space intersects the elliptic locus non-trivially. This would force the Picard rank to be \(22\) at one point and thus at every point. This idea is realized by the following steps. {\parindent=6mm \begin{itemize}\item[(1)] Using Borcherd's work, an elementary bound on coefficients of cusp forms shows the existence of an automorphic form on the moduli space of \(K3\) surfaces (over \(\mathbb{C}\)) whose zeros and poles lie in the locus of elliptic \(K3\) surfaces. This gives an ample divisor supported on the elliptic locus, which must intersect any nonisotrivial family of \(K3\) surfaces over a complete base. \item[(2)] To descend to characteristic \(p>0\), use the idea from MMP (the minimal model program) to show that any connected component of the supersingular locus contains complete curves. This requires the condition of either \(p>2d+4\) or semistable reduction. \item[(3)] The final step is to show the extension of the above ample divisor to characteristic \(p\) remains ample. This requires proving positivity of the Hodge bundle in characteristic \(p\) when \(p\geq 5\) and \(p\) does not divide \(2d\). This involves a comparison theorem in \(p\)-adic Hodge theory and the Kuga-Satake construction for \(K3\) surfaces. \end{itemize}} There are several byproducts of the above arguments. For instance, as a corollary of the third step, it is shown that the moduli space of polarized \(K3\) surfaces in characteristic \(p\) is quasi-projective when \(p\geq 5\) and \(p\) does not divide \(2d\). In the appendix by A. Snowden, a compatibility statement is proved between Clifford constructions and integral \(p\)-adic comparison functors.