Birational boundedness for holomorphic symplectic varieties, Zarhin's trick for \(K3\) surfaces, and the Tate conjecture (Q350552)

\textit{Yu. G. Zarkhin}'s trick [Math. USSR, Izv. 8, 477--480 (1975; Zbl 0332.14016); translation from Izv. Akad. Nauk SSSR, Ser. Mat. 38, 471--474 (1974)] claims that for an abelian variety \(A\) over a field \(k\), \((A\times\hat{A})^4\) admits a principal polarization. This implies the Tate conjecture for abelian varieties over finite fields [\textit{J. Tate}, Invent. Math. 2, 134--144 (1966; Zbl 0147.20303)]. The aim of the paper under review is to study a version of Zarhin's trick for \(K3\) surfaces and to give new geometric and simple proofs of the Tate conjecture for divisors of \(K3\) surfaces. The \(K3\) version of Zarhins's trick consists of two steps. The first step is to construct big line bundles on moduli spaces of sheaves on \(K3\) surfaces. More precisely, for arbitrary positive integer \(d\) there exists a positive integer \(r\) such that, for infinitely many positive integers \(m\), if \((X,H)\) is a polarized \(K3\) surface of degree \(2md\) over a field \(k\), there exists a \(4\)-dimensional smooth projective moduli space \({\mathcal M}\) of stable sheaves on \(X\) together with a line bundle \(L\) on \({\mathcal M}\) such that \(c_1(L)^4 = r\) and the Beauville-Bogomolov form \(q(K)>0\) (Theorem~1.1 and Theorem~3.3). \(L\) or its dual is big and \({\mathcal M}\) is an analogue of \((A\times \hat{A})^4\) in the original version of Zarhin's trick. The second step is to consider to what extent a birational version of Matsusaka's big theorem holds in this setting. The author proved that for arbitrary positive integers \(n\) and \(r\), the irreducible holomorphic sympletic varieties over \({\mathbb C}\) of dimension \(2n\) having the line bundles \(L\) with \(c_1(L)^{2n}=r\) and the Beauville-Bogomolov form \(q(L)>0\) are birationally bounded. Namely, there exists a projective morphism \({\mathcal X}\to S\) with a scheme \(S\) of finite type over \({\mathbb C}\) such that, for a complex point \(s\in S\), \({\mathcal X}_s\) is birational to such holomorphic symplatic variety \(X\) (Theorem~3.3 and Theorem~1.2). A weak version of this result over a finite field of characteristic \(\geq 5\) is also proved with a similar method, where the Kuga-Satake construction is used instead of the period map (Proposition~3.16). As applications, the author gives new geometric and simple proofs of the Tate conjecture for \(K3 \)surfaces over fields of characteristic \(\geq 5\) or in any characteristic when the Picard number is \(\geq 2\). The latter contains the new result for characteristic~\(2\).